Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userImproved Surrogate Modeling using Machine Learning for Industrial Civil Aircraft Aerodynamics
193
0
0.0
(
0
)
Ask ChatGPT about the research
No Arabic abstract
Predicting and simulating aerodynamic fields for civil aircraft over wide flight envelopes represent a real challenge mainly due to significant numerical costs and complex flows. Surrogate models and reduced-order models help to estimate aerodynamic fields from a few well-selected simulations. However, their accuracy dramatically decreases when different physical regimes are involved. Therefore, a method of local non-intrusive reduced-order models using machine learning, called Local Decomposition Method, has been developed to mitigate this issue. This paper introduces several enhancements to this method and presents a complex application to an industrial-like three-dimensional aircraft configuration over a full flight envelope. The enhancements of the method cover several aspects: choosing the best number of models, estimating apriori errors, improving the adaptive sampling for parallel issues, and better handling the borders between local models. The application is supported by an analysis of the model behavior, with a focus on the machine learning methods and the local properties. The model achieves strong levels of accuracy, in particular with two sub-models: one for the subsonic regime and one for the transonic regime. These results highlight that local models and machine learning represent very promising solutions to deal with surrogate models for aerodynamics.
rate research
Read More
This article presents an original methodology for the prediction of steady turbulent aerodynamic fields. Due to the important computational cost of high-fidelity aerodynamic simulations, a surrogate model is employed to cope with the significant variations of several inflow conditions. Specifically, the Local Decomposition Method presented in this paper has been derived to capture nonlinear behaviors resulting from the presence of continuous and discontinuous signals. A combination of unsupervised and supervised learning algorithms is coupled with a physical criterion. It decomposes automatically the input parameter space, from a limited number of high-fidelity simulations, into subspaces. These latter correspond to different flow regimes. A measure of entropy identifies the subspace with the expected strongest non-linear behavior allowing to perform an active resampling on this low-dimensional structure. Local reduced-order models are built on each subspace using Proper Orthogonal Decomposition coupled with a multivariate interpolation tool. The methodology is assessed on the turbulent two-dimensional flow around the RAE2822 transonic airfoil. It exhibits a significant improvement in term of prediction accuracy for the Local Decomposition Method compared with the classical method of surrogate modeling for cases with different flow regimes.
The impact of boat traffic on the health of coastal ecosystems is a multi-scale process: from minutes (individual wakes) to days (tidal modulation of sediment transport), to seasons and years (traffic is seasonal). A considerable numerical effort, notwithstanding the value of a boat-by-boat numerical modeling approach, is questionable, because of the practical impossibility of specifying the exact type and navigation characteristics for every boat comprising the traffic at any given time. Here, we propose a statistical-mechanics description of the traffic using a joint probability density of the wake population in some characteristic parameter space. We attempt to answer two basic questions: (1) what is the relevant parameter space and (2) how should a numerical model be tested for a wake population? We describe the linear and nonlinear characteristics of wakes observed in the Florida Intracoastal Waters. Adopting provisionally a two-dimensional parameter space (depth- and length-based Froude numbers) we conduct numerical simulations using the open-source FUNWAVE-TVD Boussinesq model. The model performance is excellent for weakly-dispersive, completely specified wakes (e.g., the analytical linear wakes), and also for the range of Froude numbers observed in the field, or for large container ships generating relatively long waves. The model is challenged by the short waves generated by small, slow boats. However, simulations suggest that the problem is confined to the deeper water domain and linear evolution. Nonlinear wake shoaling, essential for modeling wake-induced sediment transport and wake impact on the environment, is described well.
Wind farm design primarily depends on the variability of the wind turbine wake flows to the atmospheric wind conditions, and the interaction between wakes. Physics-based models that capture the wake flow-field with high-fidelity are computationally very expensive to perform layout optimization of wind farms, and, thus, data-driven reduced order models can represent an efficient alternative for simulating wind farms. In this work, we use real-world light detection and ranging (LiDAR) measurements of wind-turbine wakes to construct predictive surrogate models using machine learning. Specifically, we first demonstrate the use of deep autoencoders to find a low-dimensional emph{latent} space that gives a computationally tractable approximation of the wake LiDAR measurements. Then, we learn the mapping between the parameter space and the (latent space) wake flow-fields using a deep neural network. Additionally, we also demonstrate the use of a probabilistic machine learning technique, namely, Gaussian process modeling, to learn the parameter-space-latent-space mapping in addition to the epistemic and aleatoric uncertainty in the data. Finally, to cope with training large datasets, we demonstrate the use of variational Gaussian process models that provide a tractable alternative to the conventional Gaussian process models for large datasets. Furthermore, we introduce the use of active learning to adaptively build and improve a conventional Gaussian process model predictive capability. Overall, we find that our approach provides accurate approximations of the wind-turbine wake flow field that can be queried at an orders-of-magnitude cheaper cost than those generated with high-fidelity physics-based simulations.
We consider the use of probabilistic neural networks for fluid flow {surrogate modeling} and data recovery. This framework is constructed by assuming that the target variables are sampled from a Gaussian distribution conditioned on the inputs. Consequently, the overall formulation sets up a procedure to predict the hyperparameters of this distribution which are then used to compute an objective function given training data. We demonstrate that this framework has the ability to provide for prediction confidence intervals based on the assumption of a probabilistic posterior, given an appropriate model architecture and adequate training data. The applicability of the present framework to cases with noisy measurements and limited observations is also assessed. To demonstrate the capabilities of this framework, we consider canonical regression problems of fluid dynamics from the viewpoint of reduced-order modeling and spatial data recovery for four canonical data sets. The examples considered in this study arise from (1) the shallow water equations, (2) a two-dimensional cylinder flow, (3) the wake of NACA0012 airfoil with a Gurney flap, and (4) the NOAA sea surface temperature data set. The present results indicate that the probabilistic neural network not only produces a machine-learning-based fluid flow {surrogate} model but also systematically quantifies the uncertainty therein to assist with model interpretability.
The pipeline optimization problem in machine learning requires simultaneous optimization of pipeline structures and parameter adaptation of their elements. Having an elegant way to express these structures can help lessen the complexity in the management and analysis of their performances together with the different choices of optimization strategies. With these issues in mind, we created the AutoMLPipeline (AMLP) toolkit which facilitates the creation and evaluation of complex machine learning pipeline structures using simple expressions. We use AMLP to find optimal pipeline signatures, datamine them, and use these datamined features to speed-up learning and prediction. We formulated a two-stage pipeline optimization with surrogate modeling in AMLP which outperforms other AutoML approaches with a 4-hour time budget in less than 5 minutes of AMLP computation time.
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
