Data-Driven Computing With Convolutional Neural Networks for Two-Phase Flows: Application to Wave-Structure Interaction

2018 ◽  
Author(s):  
X. Mao ◽  
V. Joshi ◽  
T. P. Miyanawala ◽  
Rajeev K. Jaiman
Keyword(s):  
Neural Network ◽  
Free Surface ◽  
Bluff Body ◽  
Data Driven ◽  
Model Parameters ◽  
Wave Forces ◽  
Approximation Procedure ◽  
High Fidelity ◽  
Bluff Bodies ◽  
Fluid Forces

Fluctuating wave force on a bluff body is of great significance in many offshore and marine engineering applications. We present a Convolutional Neural Network (CNN) based data-driven computing to predict the unsteady wave forces on bluff bodies due to the free-surface wave motion. For the full-order modeling and high-fidelity data generation, the air-water interface for such wave-body problems must be captured accurately for a broad range of physical and geometric parameters. Originated from the thermodynamically consistent theories, the physically motivated Allen-Cahn phase-field method has many advantages over other interface capturing techniques such as level-set and volume-of-fluid methods. The Allen-Cahn equation is solved in the mass-conservative form by imposing a Lagrange multiplier technique. While a tremendous amount of wave-body interaction data is generated in offshore engineering via both CFD simulations and experiments, the results are generally underutilized. Design space exploration and flow control of such practical scenarios are still time-consuming and expensive. An alternative to semi-analytical modeling, CNN is a class of deep neural network for solving inverse problems which is efficient in parametric data-driven computation and can use the domain knowledge. It establishes a model with arbitrarily generated model parameters, makes predictions using the model and existing input parametric settings, and adjusts the model parameters according to the error between the predictions and existing results. The computational cost of this prediction process, compared with high-fidelity CFD simulation, is significantly reduced, which makes CNN an accessible tool in design and optimization problems. In this study, CNN-based data-driven computing is utilized to predict the wave forces on bluff bodies with different geometries and distances to the free surface. The discrete convolution process with a non-linear rectification is employed to approximate the mapping between the bluff-body shape, the distance to the free-surface and the fluid forces. The wave-induced fluid forces on bluff bodies of different shapes and submergences are predicted by the trained CNN. Finally, a convergence study is performed to identify the effective hyper-parameters of the CNN such as the convolution kernel size, the number of kernels and the learning rate. Overall, the proposed CNN-based approximation procedure has a profound impact on the parametric design of bluff bodies experiencing wave loads.

A Novel Deep Learning Method for the Predictions of Current Forces on Bluff Bodies

2018 ◽  
Author(s):  
T. P. Miyanawala ◽  
Rajeev K. Jaiman
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Bluff Body ◽  
The Body ◽  
Stochastic Gradient Descent ◽  
Approximation Procedure ◽  
Bluff Bodies ◽  
Force Coefficients ◽  
Fluid Forces ◽  
Body Shapes

Unsteady separated flow behind a bluff body causes fluctuating drag and transverse forces on the body, which is of great significance in many offshore and marine engineering applications. While physical experimental and computational techniques provide valuable physics insight, they are generally time-consuming and expensive for design space exploration and flow control of such practical scenarios. We present an efficient Convolutional Neural Network (CNN) based deep-learning technique to predict the unsteady fluid forces for different bluff body shapes. The discrete convolution process with a non-linear rectification is employed to approximate the mapping between the bluff-body shape and the fluid forces. The deep neural network is fed by the Euclidean distance function as the input and the target data generated by the full-order Navier-Stokes computations for primitive bluff body shapes. The convolutional networks are iteratively trained using a stochastic gradient descent method to predict the fluid force coefficients of different geometries and the results are compared with the full-order computations. We have extended this CNN-based technique to predict the variation of force coefficients with the Reynolds number as well. Within the error threshold, the predictions based on our deep convolutional network have a speed-up nearly three orders of magnitude compared to the full-order results and consumes an insignificant fraction of computational resources. The deep CNN-based model can predict the hydrodynamic coefficients required for the well-known Lighthill’s force decomposition in almost real time which is extremely advantageous for offshore applications. Overall, the proposed CNN-based approximation procedure has a profound impact on the parametric design of bluff bodies and the feedback control of separated flows.

Dynamic Data-Driven Spatiotemporal System Behavior Prediction With Simulations and Sensor Measurement Data

2018 ◽  
Author(s):  
Xiangxue Zhao ◽  
Shapour Azarm ◽  
Balakumar Balachandran
Keyword(s):  
Dynamical System ◽  
Real Time ◽  
Measurement Data ◽  
Data Driven ◽  
Sensor Data ◽  
Model Parameters ◽  
High Fidelity ◽  
System Behavior ◽  
Sensor Measurement

Online prediction of dynamical system behavior based on a combination of simulation data and sensor measurement data has numerous applications. Examples include predicting safe flight configurations, forecasting storms and wildfire spread, estimating railway track and pipeline health conditions. In such applications, high-fidelity simulations may be used to accurately predict a system’s dynamical behavior offline (“non-real time”). However, due to the computational expense, these simulations have limited usage for online (“real-time”) prediction of a system’s behavior. To remedy this, one possible approach is to allocate a significant portion of the computational effort to obtain data through offline simulations. The obtained offline data can then be combined with online sensor measurements for online estimation of the system’s behavior with comparable accuracy as the off-line, high-fidelity simulation. The main contribution of this paper is in the construction of a fast data-driven spatiotemporal prediction framework that can be used to estimate general parametric dynamical system behavior. This is achieved through three steps. First, high-order singular value decomposition is applied to map high-dimensional offline simulation datasets into a subspace. Second, Gaussian processes are constructed to approximate model parameters in the subspace. Finally, reduced-order particle filtering is used to assimilate sparsely located sensor data to further improve the prediction. The effectiveness of the proposed approach is demonstrated through a case study. In this case study, aeroelastic response data obtained for an aircraft through simulations is integrated with measurement data obtained from a few sparsely located sensors. Through this case study, the authors show that along with dynamic enhancement of the state estimates, one can also realize a reduction in uncertainty of the estimates.

A Lattice Boltzmann Bluff Body Model for VIV Suppression

2006 ◽  
Author(s):  
Jannette B. Frandsen
Keyword(s):  
Reynolds Number ◽  
Free Surface ◽  
Lattice Boltzmann ◽  
Water Waves ◽  
Bluff Body ◽  
Numerical Models ◽  
Bluff Bodies ◽  
Free Surfaces ◽  
Viv Suppression ◽  
Reynolds Number Flows

In this paper, the suitability of a mesoscopic approach involving a single phase Lattice Boltzmann (LB) model is examined. In contrast, to continuum based numerical models, where only space and time are discrete, the discrete variables of the LB model are space, time and particle velocity. With reference to the Boltzmann equation of classical kinetic theory, the distribution of fluid molecules is represented by particle distribution functions. The LB method simulates fluid flow by tracking particle distributions. It is notable that the formulation avoids the need to include the Poisson equation. An elastic-collision scheme with no-slip walls is prescribed. The central idea behind proposing the present formulation is many fold. One goal is to capture smaller scales naturally, postponing the need of applying empirical turbulence models. Another goal is to get further insight into nonlinearities in steep and breaking free surfaces to improve current continuum mechanics solutions. Although the long term goal is to predict bluff-body high Reynolds number flows and breaking water waves, the present study is limited to laminar flow simulations and continuous free surfaces. The case studies presented include bluff bodies embedded in Reynolds number flows in the order of 100–200. The free surface test cases represent bore propagation past a single and multiple structures. The 2-D uniform grid solutions are compared with findings reported in the literature. Vortex patterns are studied when single or several objects are located in the bluff-body wakes. From a mitigation point of view, the model presents an easy means of re-arranging bluff bodies to study optimum solutions for VIV suppression with/without a free surface.

scholarly journals Reaction Wheels Fault Isolation Onboard 3-Axis Controlled Satel-lite using Enhanced Random Forest with Multidomain Features

2021 ◽  
Vol 12 (2) ◽  
Author(s):  
Afshin Rahimi ◽  
Mofiyinoluwa O. Folami
Keyword(s):  
Machine Learning ◽  
Fault Isolation ◽  
Data Driven ◽  
Machine Learning Techniques ◽  
Superior Performance ◽  
Alternative Methods ◽  
Model Parameters ◽  
High Fidelity ◽  
Novel Approach ◽  
Learning Techniques

As the number of satellite launches increases each year, it is only natural that an interest in the safety and monitoring of these systems would increase as well. However, as a system becomes more complex, generating a high-fidelity model that accurately describes the system becomes complicated. Therefore, imploring a data-driven method can provide to be more beneficial for such applications. This research proposes a novel approach for data-driven machine learning techniques on the detection and isolation of nonlinear systems, with a case-study for an in-orbit closed loop-controlled satellite with reaction wheels as actuators. High-fidelity models of the 3-axis controlled satellite are employed to generate data for both nominal and faulty conditions of the reaction wheels. The generated simulation data is used as input for the isolation method, after which the data is pre-processed through feature extraction from a temporal, statistical, and spectral domain. The pre-processed features are then fed into various machine learning classifiers. Isolation results are validated with cross-validation, and model parameters are tuned using hyperparameter optimization. To validate the robustness of the proposed method, it is tested on three characterized datasets and three reaction wheel configurations, including standard four-wheel, three-orthogonal, and pyramid. The results prove superior performance isolation accuracy for the system under study compared to previous studies using alternative methods (Rahimi & Saadat, 2019, 2020).

A Data-Driven Approach for the Stability Analysis of Vortex-Induced Vibration

2018 ◽  
Author(s):  
Sandeep B. Reddy ◽  
Allan Ross Magee ◽  
Rajeev K. Jaiman
Keyword(s):  
Stability Analysis ◽  
Bluff Body ◽  
Data Driven ◽  
Bluff Bodies ◽  
Stability Boundaries ◽  
Data Driven Approach ◽  
Offshore Industry ◽  
The Stability ◽  
Lock In ◽  
And Control

In this paper, a general data-driven approach to construct a reduced-order model (ROM) for the coupled fluid-structure interaction (FSI) problem of a transversely vibrating bluff body in an incompressible flow is presented. The proposed data-driven approach relies on the Eigensystem Realization Algorithm (ERA) to design ROM models in a state-space format. The stability boundaries of the coupled FSI system are obtained by examining the eigenvalue trajectories of the ERA-based ROM. These stability boundaries provide us valuable quantitative insights into the lock-in phenomenon of the bluff-body vibration. We demonstrate the present ERA-based ROM technique for various configurations of bluff bodies such as an isolated single cylinder, the side-by-side and the tandem cylinder arrangements. A comparative study on the effect of different appendages to suppress the VIV of a cylinder is also presented using the ERA-based stability analysis. The validity of the proposed method for the FSI stability analysis on such variety of configurations has not been presented before and is the novel contribution of this paper. Overall, the proposed data-driven framework is found to be much more effective in terms of computational costs and the predicted lock-in regions are comparable to high-fidelity full-order simulations. This work has a potential for a profound impact on the design optimization and control of bluff body structures used in offshore industry.

scholarly journals Hybrid Modelling by Machine Learning Corrections of Analytical Model Predictions towards High-Fidelity Simulation Solutions

Materials ◽  
2021 ◽  
Vol 14 (8) ◽  
pp. 1883
Author(s):  
Frederic E. Bock ◽  
Sören Keller ◽  
Norbert Huber ◽  
Benjamin Klusemann
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Analytical Model ◽  
Data Driven ◽  
High Fidelity ◽  
Prediction Errors ◽  
Hybrid Modelling ◽  
Model Predictions ◽  
Physical Laws ◽  
Modelling Approach

Within the fields of materials mechanics, the consideration of physical laws in machine learning predictions besides the use of data can enable low prediction errors and robustness as opposed to predictions only based on data. On the one hand, exclusive utilization of fundamental physical relationships might show significant deviations in their predictions compared to reality, due to simplifications and assumptions. On the other hand, using only data and neglecting well-established physical laws can create the need for unreasonably large data sets that are required to exhibit low bias and are usually expensive to collect. However, fundamental but simplified physics in combination with a corrective model that compensates for possible deviations, e.g., to experimental data, can lead to physics-based predictions with low prediction errors, also despite scarce data. In this article, it is demonstrated that a hybrid model approach consisting of a physics-based model that is corrected via an artificial neural network represents an efficient prediction tool as opposed to a purely data-driven model. In particular, a semi-analytical model serves as an efficient low-fidelity model with noticeable prediction errors outside its calibration domain. An artificial neural network is used to correct the semi-analytical solution towards a desired reference solution provided by high-fidelity finite element simulations, while the efficiency of the semi-analytical model is maintained and the applicability range enhanced. We utilize residual stresses that are induced by laser shock peening as a use-case example. In addition, it is shown that non-unique relationships between model inputs and outputs lead to high prediction errors and the identification of salient input features via dimensionality analysis is highly beneficial to achieve low prediction errors. In a generalization task, predictions are also outside the process parameter space of the training region while remaining in the trained range of corrections. The corrective model predictions show substantially smaller errors than purely data-driven model predictions, which illustrates one of the benefits of the hybrid modelling approach. Ultimately, when the amount of samples in the data set is reduced, the generalization of the physics-related corrective model outperforms the purely data-driven model, which also demonstrates efficient applicability of the proposed hybrid modelling approach to problems where data is scarce.

scholarly journals Data-Driven Stability Assessment of Multilayer Long Short-Term Memory Networks

2021 ◽  
Vol 11 (4) ◽  
pp. 1829
Author(s):  
Davide Grande ◽  
Catherine A. Harris ◽  
Giles Thomas ◽  
Enrico Anderlini
Keyword(s):  
Neural Network ◽  
Network Architecture ◽  
Short Term Memory ◽  
Moving Average ◽  
Machine Learning Algorithms ◽  
Physical Models ◽  
Data Driven ◽  
The Stability ◽  
And Control ◽  
Nonlinear Autoregressive

Recurrent Neural Networks (RNNs) are increasingly being used for model identification, forecasting and control. When identifying physical models with unknown mathematical knowledge of the system, Nonlinear AutoRegressive models with eXogenous inputs (NARX) or Nonlinear AutoRegressive Moving-Average models with eXogenous inputs (NARMAX) methods are typically used. In the context of data-driven control, machine learning algorithms are proven to have comparable performances to advanced control techniques, but lack the properties of the traditional stability theory. This paper illustrates a method to prove a posteriori the stability of a generic neural network, showing its application to the state-of-the-art RNN architecture. The presented method relies on identifying the poles associated with the network designed starting from the input/output data. Providing a framework to guarantee the stability of any neural network architecture combined with the generalisability properties and applicability to different fields can significantly broaden their use in dynamic systems modelling and control.

scholarly journals Data-driven RANS closures for three-dimensional flows around bluff bodies

2021 ◽  
pp. 104997
Author(s):  
Jasper P. Huijing ◽  
Richard P. Dwight ◽  
Martin Schmelzer
Keyword(s):  
Three Dimensional ◽  
Data Driven ◽  
Bluff Bodies
Sign in / Sign up

Export Citation Format

Share Document