SEARCH · Search NASA

Results for “gradient based optimization”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 469 records · Page 26

All sources NASA NTRS DOE OSTI

DOE OSTI2020

A Hybrid Gradient Method to Designing Bayesian Experiments for Implicit Models

Bayesian experimental design (BED) aims at designing an experiment to maximize the information gathering from the collected data. The optimal design is usually achieved by maximizing the mutual information (MI) between the data and the model parameters. When the analytical expression of the MI is unavailable, e.g.,having implicit models with intractable data distributions, a neural network-based lower bound of the MI was recently proposed and a gradient ascent method was used to maximize the lower bound [1]. However, the approach in [1] requires a pathwise sampling path to compute the gradient of the MI lower bound with respect to the design variables, and such a pathwise sampling path is usually inaccessible for implicit models. In this work, we propose a hybrid gradient approach that leverages recent advances in variational MI estimator and evolution strategies (ES)combined with black-box stochastic gradient ascent (SGA) to maximize the MI lower bound. This allows the design process to be achieved through a unified scalable procedure for implicit models without sampling path gradients. Several experiments demonstrate that our approach significantly improves the scalability of BED for implicit models in high-dimensional design space.

Zhang, Jiaxin↗

NASA NTRS1996

Aerodynamic Shape Optimization of Supersonic Aircraft Configurations via an Adjoint Formulation on Parallel Computers

This work describes the application of a control theory-based aerodynamic shape optimization method to the problem of supersonic aircraft design. The design process is greatly accelerated through the use of both control theory and a parallel implementation on distributed memory computers. Control theory is employed to derive the adjoint differential equations whose solution allows for the evaluation of design gradient information at a fraction of the computational cost required by previous design methods. The resulting problem is then implemented on parallel distributed memory architectures using a domain decomposition approach, an optimized communication schedule, and the MPI (Message Passing Interface) Standard for portability and efficiency. The final result achieves very rapid aerodynamic design based on higher order computational fluid dynamics methods (CFD). In our earlier studies, the serial implementation of this design method was shown to be effective for the optimization of airfoils, wings, wing-bodies, and complex aircraft configurations using both the potential equation and the Euler equations. In our most recent paper, the Euler method was extended to treat complete aircraft configurations via a new multiblock implementation. Furthermore, during the same conference, we also presented preliminary results demonstrating that this basic methodology could be ported to distributed memory parallel computing architectures. In this paper, our concern will be to demonstrate that the combined power of these new technologies can be used routinely in an industrial design environment by applying it to the case study of the design of typical supersonic transport configurations. A particular difficulty of this test case is posed by the propulsion/airframe integration.

Reuther, James↗

DOE OSTI2024

Optimal control of large quantum systems: assessing memory and runtime performance of GRAPE

Abstract Gradient Ascent Pulse Engineering (GRAPE) is a popular technique in quantum optimal control, and can be combined with automatic differentiation (AD) to facilitate on-the-fly evaluation of cost-function gradients. We illustrate that the convenience of AD comes at a significant memory cost due to the cumulative storage of a large number of states and propagators. For quantum systems of increasing Hilbert space size, this imposes a significant bottleneck. We revisit the strategy of hard-coding gradients in a scheme that fully avoids propagator storage and significantly reduces memory requirements. Separately, we present improvements to numerical state propagation to enhance runtime performance. We benchmark runtime and memory usage and compare this approach to AD-based implementations, with a focus on pushing towards larger Hilbert space sizes. The results confirm that the AD-free approach facilitates the application of optimal control for large quantum systems which would otherwise be difficult to tackle.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2020

Eigenfrequency constrained topology optimization of finite strain hyperelastic structures

This paper incorporates hyperelastic materials, nonlinear kinematics, and preloads in eigenfrequency constrained density–based topology optimization. The formulation allows for initial finite deformations and subsequent small harmonic oscillations. The optimization problem is solved by the method of moving asymptotes, and the gradients are calculated using the adjoint method. Both simple and degenerate eigenfrequencies are considered in the sensitivity analysis. A well-posed topology optimization problem is formulated by filtering the volume fraction field. Numerical issues associated with excessive distortion and spurious eigenmodes in void regions are reduced by removing low volume fraction elements. The optimization objective is to maximize stiffness subject to a lower bound on the fundamental eigenfrequency. Numerical examples show that the eigenfrequencies drastically change with the load magnitude, and that the optimization is able to produce designs with the desired fundamental eigenfrequency.

42 ENGINEERING↗

DOE OSTI2026

A Machine Learning based Approach of Estimating Equivalent Circuit Model Parameters at Different SoCs of Li-ion Batteries from Voltage Relaxation

Abstract: In this study, an approach of estimating the equivalent circuit model (ECM) parameters for Li-ion batteries (LIBs) is proposed based on the voltage value at different intervals while relaxing the LIB after discharge. The typical approach for estimating ECM parameters of a LIB is to conduct electrochemical impedance spectroscopy (EIS) measurements at different frequencies and fit them to a predefined circuit model, which requires additional measuring arrangements and specialized devices. The proposed methodology utilizes four different voltages at 0s, 60s, 360s, and 1800s alongside the specific state of charge (SoC) value for a specific constant discharge current value of ~1C until the relaxation stage to train and evaluate three regression-based machine learning models— Support Vector Regression (SVR), Extreme Gradient Boosting (XGBoost), and Gaussian Process Regression (GPR)—for estimating the ECM parameters of the selected model. Bayesian optimization is employed for hyperparameter tuning to achieve optimal performance for all the regressor models, among which, the GPR provided the best performance with the root-mean-squared error (RMSE) of less than 4x10-4 on average for the resistive components and less than 0.27 for capacitive components with excellent R2 scores. The simplicity of the approach enables it to eliminate the need for sophisticated measuring equipment and computation power.

Sagar, Md. Samiul [The University of Alabama (UA)]↗

NASA NTRS1986

Optimization and gamma/theta guidance of flight trajectories in a windshear

Guidance schemes for near-optimum performance in wind shear are examined. The presence of a downdraft is assumed in addition to the horizontal shear. The takeoff problem is considered with reference to flight in a vertical plane. Trajectories for optimum performance in a wind shear are determined for different wind shear models and intensities. Methods of optimal control theory are used together with the dual sequential gradient-restoration algorithm for optimal control problems. Guidance schemes for near-optimum performance in a wind shear are developed, starting from optimal trajectories. These are gamma guidance, based on either the absolute of the relative path inclination, and theta guidance, based on the pitch attitude angle. These schemes are evaluated through numerical experiments in order to determine whether the resulting trajectories are sufficiently close to optimum and to compare these trajectories with those arising from alternative guidance schemes.

Miele, A.↗

NASA NTRS1992

A conjugate gradients/trust regions algorithms for training multilayer perceptrons for nonlinear mapping

This paper addresses the issue of applying a globally convergent optimization algorithm to the training of multilayer perceptrons, a class of Artificial Neural Networks. The multilayer perceptrons are trained towards the solution of two highly nonlinear problems: (1) signal detection in a multi-user communication network, and (2) solving the inverse kinematics for a robotic manipulator. The research is motivated by the fact that a multilayer perceptron is theoretically capable of approximating any nonlinear function to within a specified accuracy. The algorithm that has been employed in this study combines the merits of two well known optimization algorithms, the Conjugate Gradients and the Trust Regions Algorithms. The performance is compared to a widely used algorithm, the Backpropagation Algorithm, that is basically a gradient-based algorithm, and hence, slow in converging. The performances of the two algorithms are compared with the convergence rate. Furthermore, in the case of the signal detection problem, performances are also benchmarked by the decision boundaries drawn as well as the probability of error obtained in either case.

Madyastha, Raghavendra K.↗

NASA NTRS1996

Aerodynamic Shape Optimization of Supersonic Aircraft Configurations via an Adjoint Formulation on Parallel Computers

This work describes the application of a control theory-based aerodynamic shape optimization method to the problem of supersonic aircraft design. The design process is greatly accelerated through the use of both control theory and a parallel implementation on distributed memory computers. Control theory is employed to derive the adjoint differential equations whose solution allows for the evaluation of design gradient information at a fraction of the computational cost required by previous design methods (13, 12, 44, 38). The resulting problem is then implemented on parallel distributed memory architectures using a domain decomposition approach, an optimized communication schedule, and the MPI (Message Passing Interface) Standard for portability and efficiency. The final result achieves very rapid aerodynamic design based on higher order computational fluid dynamics methods (CFD). In our earlier studies, the serial implementation of this design method (19, 20, 21, 23, 39, 25, 40, 41, 42, 43, 9) was shown to be effective for the optimization of airfoils, wings, wing-bodies, and complex aircraft configurations using both the potential equation and the Euler equations (39, 25). In our most recent paper, the Euler method was extended to treat complete aircraft configurations via a new multiblock implementation. Furthermore, during the same conference, we also presented preliminary results demonstrating that the basic methodology could be ported to distributed memory parallel computing architectures [241. In this paper, our concem will be to demonstrate that the combined power of these new technologies can be used routinely in an industrial design environment by applying it to the case study of the design of typical supersonic transport configurations. A particular difficulty of this test case is posed by the propulsion/airframe integration.

Reuther, James↗

NASA NTRS1984

A rotor optimization using regression analysis

The design and development of helicopter rotors is subject to the many design variables and their interactions that effect rotor operation. Until recently, selection of rotor design variables to achieve specified rotor operational qualities has been a costly, time consuming, repetitive task. For the past several years, Kaman Aerospace Corporation has successfully applied multiple linear regression analysis, coupled with optimization and sensitivity procedures, in the analytical design of rotor systems. It is concluded that approximating equations can be developed rapidly for a multiplicity of objective and constraint functions and optimizations can be performed in a rapid and cost effective manner; the number and/or range of design variables can be increased by expanding the data base and developing approximating functions to reflect the expanded design space; the order of the approximating equations can be expanded easily to improve correlation between analyzer results and the approximating equations; gradients of the approximating equations can be calculated easily and these gradients are smooth functions reducing the risk of numerical problems in the optimization; the use of approximating functions allows the problem to be started easily and rapidly from various initial designs to enhance the probability of finding a global optimum; and the approximating equations are independent of the analysis or optimization codes used.

Giansante, N.↗

DOE OSTI2026

Physicochemical and Performance Characterization of Six Commercial Organic Solvent Nanofiltration Membranes

This work introduces a novel, gradient-free metamaterial design method based on Gaussian process regression to represent the density field of a unit cell. The dimension of the design space is determined by the covariance matrix dimension in the Gaussian process regression. We propose compressing this matrix using an autoencoder, enabling the decoder to generate the density field and effectively reduce the originally large design space to a lower-dimensional subspace. In this compressed space, we employ an active learning method, Bayesian Adaptive Direct Search (BADS), for efficient exploration of the design space. We demonstrate that for simple 2D designs aimed at maximizing unit cell stiffness, our method yields results comparable to those of standard topology optimization. Furthermore, we extend our approach to various mechanical problems, from linear elasticity to hyperelastic large deformation and elasto-plasticity under finite deformation, to 3D metamaterial design. This illustrates the method’s versatility and effectiveness across a range of applications.

Wu, Haoran↗

NASA NTRS1991

Aerodynamic design optimization using sensitivity analysis and computational fluid dynamics

A new and efficient method is presented for aerodynamic design optimization, which is based on a computational fluid dynamics (CFD)-sensitivity analysis algorithm. The method is applied to design a scramjet-afterbody configuration for an optimized axial thrust. The Euler equations are solved for the inviscid analysis of the flow, which in turn provides the objective function and the constraints. The CFD analysis is then coupled with the optimization procedure that uses a constrained minimization method. The sensitivity coefficients, i.e. gradients of the objective function and the constraints, needed for the optimization are obtained using a quasi-analytical method rather than the traditional brute force method of finite difference approximations. During the one-dimensional search of the optimization procedure, an approximate flow analysis (predicted flow) based on a first-order Taylor series expansion is used to reduce the computational cost. Finally, the sensitivity of the optimum objective function to various design parameters, which are kept constant during the optimization, is computed to predict new optimum solutions. The flow analysis of the demonstrative example are compared with the experimental data. It is shown that the method is more efficient than the traditional methods.

Baysal, Oktay↗

DOE OSTI2022

Optimal Coordination of Distributed Energy Resources Using Deep Deterministic Policy Gradient

Recent studies showed that reinforcement learning (RL) is a promising approach for coordination and control of distributed energy resources (DER) under uncertainties. Many existing RL approaches, including Q-learning and approximate dynamic programming, are based on lookup table methods, which become inefficient when the problem size is large and infeasible when continuous states and actions are involved. In addition, when modeling battery energy storage system (BESS), the loss of life is not reasonably considered into the decision-making process. This paper proposes an innovative deep RL method for DER coordination considering BESS degradation. The proposed deep RL is designed based on an adaptive actor-critic architecture and employs an off-policy deterministic policy gradient method for determining the dispatch operation that minimizes the operation cost and BESS life loss. Case studies were performed to validate the proposed method and demonstrate the effects of incorporating degradation models into control design.

Das, Avijit↗

DOE OSTI2023

Gradient process parameter optimization in additive friction stir deposition of aluminum alloys

As one of the most novel additive manufacturing methods, currently selection and optimization of processing parameters in additive friction stir deposition (AFSD) have mainly relied on experiments and subsequent characterization of microstructural and mechanical properties. Such approaches are both time- and resource-consuming. Therefore, an ultrasound elastography enhanced gradient process parameter optimization method was applied in the present work to obtain a window of optimized processing parameters for AFSD processing of aluminum alloy by varying both rotational and linear deposition speeds. Further, the quality of AFSD processed layer was investigated for physical nature of surface, dynamic elastic modulus, and microstructural aspects in cross-sections of the deposited layer. The efficiency in exploring process parameters was significantly enhanced by implementing a high-throughput screening experimental design based on application of gradient process parameters and continuous ultrasound elastographs. In addition, the applied ultrasonic elastography technique assisted in evaluating the homogeneity in microstructure and mechanical properties of AFSD sample over the entire gradients of the process parameters. The techniques adopted in current work can be further extended to identify suitable parameters for AFSD fabrication of components with desired mechanical properties such as hardness, fatigue, etc.

36 MATERIALS SCIENCE↗

DOE OSTI2020

Multimodal Data Fusion via Entropy Minimization

The use of gradient-based data-driven models to solve a range of real-world remote sensing problems can in practice be limited by the uniformity of available data. Use of data from disparate sensor types, resolutions, and qualities typically requires compromises based on assumptions that are made prior to model training and may not necessarily be optimal given over-arching objectives. For example, while deep neural networks (NNs) are state-of-the-art in a variety of target detection problems, training them typically requires either limiting the training data to a subset over which uniformity can be enforced or training independent models which subsequently require additional score fusion. The method we introduce here seeks to leverage the benefits of both approaches by allowing correlated inputs from different data sources to co-influence preferred model solutions, while maintaining flexibility over missing and mismatching data. In this work we propose a new data fusion technique for gradient updated models based on entropy minimization and experimentally validate it on a hyperspectral target detection dataset. We demonstrate superior performance compared to currently available techniques using a range of realistic data scenarios, where available data has limited spacial overlap and resolution.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS1997

Crystal Growth Control

We present an innovative design of a vertical transparent multizone furnace which can operate in the temperature range of 25 C to 750 C and deliver thermal gradients of 2 C/cm to 45 C/cm for the commercial applications to crystal growth. The operation of the eight zone furnace is based on a self-tuning temperature control system with a DC power supply for optimal thermal stability. We show that the desired thermal profile over the entire length of the furnace consists of a functional combination of the fundamental thermal profiles for each individual zone obtained by setting the set-point temperature for that zone. The self-tuning system accounts for the zone to zone thermal interactions. The control system operates such that the thermal profile is maintained under thermal load, thus boundary conditions on crystal growth ampoules can be predetermined prior to crystal growth. Temperature profiles for the growth of crystals via directional solidification, vapor transport techniques, and multiple gradient applications are shown to be easily implemented. The unique feature of its transparency and ease of programming thermal profiles make the furnace useful for scientific and commercial applications for the determination of process parameters to optimize crystal growth conditions.

Duval, Walter M. B.↗

NASA NTRS1998

The Design of a Transparent Vertical Multizone Furnace: Application to Thermal Field Tuning and Crystal Growth

We present an innovative design of a vertical transparent multizone furnace which can operate in the temperature range of 25 C to 750 C and deliver thermal gradients of 2 C/cm to 45 C/cm for the commercial applications to crystal growth. The operation of the eight zone furnace is based on a self-tuning temperature control system with a DC power supply for optimal thermal stability. We show that the desired thermal profile over the entire length of the furnace consists of a functional combination of the fundamental thermal profiles for each individual zone obtained by setting the set-point temperature for that zone. The self-tuning system accounts for the zone to zone thermal interactions. The control system operates such that the thermal profile is maintained under thermal load, thus boundary conditions on crystal growth ampoules can be predetermined prior to crystal growth. Temperature profiles for the growth of crystals via directional solidification, vapor transport techniques, and multiple gradient applications are shown to be easily implemented. The unique feature of its transparency and ease of programming thermal profiles make the furnace useful in scientific and commercial applications for determining the optimized process parameters for crystal growth.

Duvual, Walter M. B.↗

DOE OSTI2024

Learning Constrained Parametric Differentiable Predictive Control Policies With Guarantees

We present differentiable predictive control (DPC), a method for offline learning of constrained neural control policies for nonlinear dynamical systems with performance guarantees. We show that the sensitivities of the parametric optimal control problem can be used to obtain direct policy gradients. Specifically, we employ automatic differentiation (AD) to efficiently compute the sensitivities of the model predictive control (MPC) objective function and constraints penalties. To guarantee safety upon deployment, we derive probabilistic guarantees on closed-loop stability and constraint satisfaction based on indicator functions and Hoeffding’s inequality. We empirically demonstrate that the proposed method can learn neural control policies for various parametric optimal control tasks. In particular, we show that the proposed DPC method can stabilize systems with unstable dynamics, track time-varying references, and satisfy nonlinear state and input constraints. Our DPC method has practical time savings compared to alternative approaches for fast and memory-efficient controller design. Specifically, DPC does not depend on a supervisory controller as opposed to approximate MPC based on imitation learning. We demonstrate that, without losing performance, DPC is scalable with greatly reduced demands on memory and computation compared to implicit and explicit MPC while being more sample efficient than model-free reinforcement learning (RL) algorithms.

97 MATHEMATICS AND COMPUTING↗

NASA NTRS2003

Optimization of single-base-pair mismatch discrimination in oligonucleotide microarrays

The discrimination between perfect-match and single-base-pair-mismatched nucleic acid duplexes was investigated by using oligonucleotide DNA microarrays and nonequilibrium dissociation rates (melting profiles). DNA and RNA versions of two synthetic targets corresponding to the 16S rRNA sequences of Staphylococcus epidermidis (38 nucleotides) and Nitrosomonas eutropha (39 nucleotides) were hybridized to perfect-match probes (18-mer and 19-mer) and to a set of probes having all possible single-base-pair mismatches. The melting profiles of all probe-target duplexes were determined in parallel by using an imposed temperature step gradient. We derived an optimum wash temperature for each probe and target by using a simple formula to calculate a discrimination index for each temperature of the step gradient. This optimum corresponded to the output of an independent analysis using a customized neural network program. These results together provide an experimental and analytical framework for optimizing mismatch discrimination among all probes on a DNA microarray.

NASA Discipline Life Sciences Technologies↗