Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Linear 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.

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

At least 1,513 records · Page 84

Quasi-Stochastic Approximation and Off-Policy Reinforcement Learning: Preprint

The Robbins-Monro stochastic approximation algorithm is a foundation of many algorithmic frameworks for reinforcement learning and often an efficient approach to solving (or approximating the solution to) complex optimal control problems. In many cases, however, practitioners are unable to apply these techniques because of an inherent high variance. This paper aims to provide a general foundation for 'quasi-stochastic approximation,' in which all of the processes under consideration are deterministic, much like quasi-Monte Carlo for variance reduction in simulation. The variance reduction can be substantial, subject to tuning of pertinent parameters in the algorithm. This paper introduces a new coupling argument to establish the optimal rate of convergence provided the gain is sufficiently large. These results are established for linear models and tested also in nonideal settings. A major application of these general results is a new class of reinforcement learning algorithms for deterministic state space models. In this setting, the main contribution is a class of algorithms for approximating the value function for a given policy, using a different policy designed to introduce exploration.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Probabilistic Forecast of All-Sky Solar Radiation Using Enhanced WRF-Solar: Preprint

This study presents enhancements of the Weather Research and Forecasting model with solar extensions (WRF-Solar) to provide probabilistic forecasts of solar radiation. Our approach builds ensemble WRF-Solar runs by introducing stochastic perturbations of variables that produce the largest uncertainties in predicting surface irradiance and clouds. The key variables are identified using tangent linear sensitivity analysis of six physics packages responsible for all-sky irradiance variability. An optimal strategy to stochastically perturb the selected variables is developed and applied to WRF-Solar to generate ensemble members for day-ahead solar prediction. The National Solar Radiation Database (NSRDB) is used to validate the ensemble forecast at arbitrary locations on the model grid. Preliminary results indicate that the proposed technique can potentially produce WRF-Solar ensembles providing reliable information of solar prediction uncertainty. This study describes the implemented methodology and initial results as well as future research to improve ensemble-based probabilistic forecasts with WRF-Solar.

all-sky↗

Thermal Analysis for SRS Saltstone Disposal Unit Facility

Saltstone Disposal Units (SDU) are long term disposal units for low activity grout produced from the solidification of decontaminated non-hazardous salt waste at SRS. The unit is a cylindrical disposal vault of 150 ft. in diameter and 23 ft. in height. The total volume the SDU can hold is 3 million gallons of grout. The hydration process of the grout material produces heat from an exothermic chemical reaction. If the grout reaches temperatures higher than the maximum safety temperature of 95 deg. C, cracks can start forming leading to contamination of the environment. The purpose of this study is to develop the optimal baseline domain boundary conditions and to perform a sensitivity analysis with respect to the baseline modeling conditions. The following assumptions were made for the modeling calculations: - The system geometry is axisymmetric; - The system was at steady state; - Ambient temperature was set to 100 deg. F; - Air was assumed to be an incompressible ideal gas; - Radiation was neglected for a conservative estimation; - Natural convection for the SDU vapor space is assumed to be turbulent because Rayleigh number based on typical operating conditions is higher than 10{sup 8}; - Water tables were not taken into consideration. The following sensitivity analyses were performed to find the optimal domain boundary: - Mesh size sensitivity; - Soil depth sensitivity - Soil depth varied from 30 ft to 250 ft with a soil thickness of 90 ft; - Soil thickness sensitivity - Soil thickness varied from 30 ft to 270 ft with a soil depth of 180 ft; - Soil conductivity sensitivity - Soil thermal conductivity varied from 0.25 W/m-K to 1 W/m-K; - Ambient Temperature Sensitivity - Ambient temperature varied from 70 deg. F to 110 deg. F. For the thermal calculations, the temperature of the bottom of the soil was set to 17.4 deg. C. The heat flux on the roof of the building and surface of the earth is 163.2 W/m{sup 2}. The soil sides are set to be adiabatic. The sensitivity analysis shows that the mesh size has no effect on the numerical results. This is mainly due to a small convection zone and large conduction zones in the modeling domain. The optimal domain boundary was found to be a soil thickness of 90 ft and a soil depth of 150 ft. The different sensitivity analysis performed here shows the optimal boundary domain to be used for thermal calculations is a soil depth of 150 ft beneath the SDU and a soil thickness of 90 ft. Anything larger would produce similar solutions but at a longer computational time to solve. The maximum temperature was more sensitive to soil properties as shown here. While the ambient temperature affected the max temperature linearly. Most of the heat transfer takes place downward vertically between the SDU and the soil beneath it. Using the optimal domain boundary determined by the sensitivity analyses, the maximum thermal loading to keep the maximum grout temperature below 95 deg. C is 1.646 W/m{sup 3}.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

Laser-Induced Spectrochemical Assay for Uranium Enrichment (LISA-UE)

Uranium hexafluoride (UF6) is the uranium compound typically involved in uranium enrichment process. As the first line of defense against nuclear proliferation, accurate determinations of the uranium enrichment ratio in UF6 are critical for materials verification, accounting and safeguards. Shipping gaseous UF6 samples off-site for analysis with mass spectrometry is cumbersome and costly, and results are not available for some time (months). In-field UF6 enrichment assay has the potential to substantially reduce the time, logistics and expense of sample handling. At present, COMPUCEA is the only accepted method for UF6 enrichment assay in the field. Laser-Induced Spectrochemical Assay for Uranium Enrichment (LISA-UE) is an all-optical (based on laser induced plasma emission) analytical technique intended for fieldable, accurate, precise and rapid UF6 enrichment assay. In its operation, laser induced plasma is created directly in the gaseous UF6 sample. Because different U isotopes emit at slightly different wavelengths, the isotopic information of the UF6 sample is inherently encoded in the atomic emission from the plasma. Isotopic emissions from 235U and 238U are measured simultaneously, which eliminate correlated noise from the laser induced plasma. Isotopic information of the UF6 sample can be extracted from the acquired spectrum with theoretical multi-variable non-linear spectral fitting. To date, advances made by the LISA-UE research team include optimization of the spectral window for direct gaseous UF6 enrichment assay with laser induced plasma, development of data reduction algorithms, and demonstrations of the LISA-UE technique with gaseous UF6 samples. In this presentation, the technical aspect of LISA-UE will be overviewed, the data reduction algorithm will be described, and performance of the technique will be discussed.

Chan, George↗

Towards an optimal reconstruction of baryon oscillations

The Baryon Acoustic Oscillations (BAO) in the large-scale structure of the universe leave a distinct peak in the two-point correlation function of the matter distribution. That acoustic peak is smeared and shifted by bulk flows and non-linear evolution. However, it has been shown that it is still possible to sharpen the peak and remove its shift by undoing the effects of the bulk flows. We propose an improvement to the standard acoustic peak reconstruction. Contrary to the standard approach, the new scheme has no free parameters, treats the large-scale modes consistently, and uses optimal filters to extract the BAO information. At redshift of zero, the reconstructed linear matter power spectrum leads to a markedly improved sharpening of the reconstructed acoustic peak compared to standard reconstruction.

79 ASTRONOMY AND ASTROPHYSICS↗

Multiresolution GPC-Structured Control of a Single-Loop Cold-Flow Chemical Looping Testbed

Chemical looping is a near-zero emission process for generating power from coal. It is based on a multi-phase gas-solid flow and has extremely challenging nonlinear, multi-scale dynamics with jumps, producing large dynamic model uncertainty, which renders traditional robust control techniques, such as linear parameter varying H ∞ design, largely inapplicable. This process complexity is addressed in the present work through the temporal and the spatiotemporal multiresolution modeling along with the corresponding model-based control laws. Namely, the nonlinear autoregressive with exogenous input model structure, nonlinear in the wavelet basis, but linear in parameters, is used to identify the dominant temporal chemical looping process dynamics. The control inputs and the wavelet model parameters are calculated by optimizing a quadratic cost function using a gradient descent method. The respective identification and tracking error convergence of the proposed self-tuning identification and control schemes, the latter using the unconstrained generalized predictive control structure, is separately ascertained through the Lyapunov stability theorem. The rate constraint on the control signal in the temporal control law is then imposed and the control topology is augmented by an additional control loop with self-tuning deadbeat controller which uses the spatiotemporal wavelet riser dynamics representation. The novelty of this work is three-fold: (1) developing the self-tuning controller design methodology that consists in embedding the real-time tunable temporal highly nonlinear, but linearly parametrizable, multiresolution system representations into the classical rate-constrained generalized predictive quadratic optimal control structure, (2) augmenting the temporal multiresolution loop by a more complex spatiotemporal multiresolution self-tuning deadbeat control loop, and (3) demonstrating the effectiveness of the proposed methodology in producing fast recursive real-time algorithms for controlling highly uncertain nonlinear multiscale processes. The latter is shown through the data from the implemented temporal and augmented spatiotemporal solutions of a difficult chemical looping cold flow tracking control problem.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

PowerMappeR: Power-Optimized Mapping of SNNs onto ReRAM Crossbars coupled via Packet-Switched NoCs

Many recent efforts in developing hardware-accelerated spiking neural networks (SNNs) are characterized by deep co-design between algorithms, architectures, and devices. Architectural advances overcome device constraints by coupling together many small resistive-RAM (ReRAM) crossbars via a network-on-chip (NoC) for neuromorphic component operation. Concurrently, improved SNN training methods increase accuracy and structural sparsity in networks despite growing problem sizes. Finally, compilers leverage these attributes to minimize area and inter-crossbar communication while mapping large SNNs to sophisticated architectures. However, for compiler-driven co-design to realize increasingly complex and profitable optimizations, a compile-time view of power consumption is critical. We present PowerMappeR to express and optimize over mapping-, architecture-, and device-specific power consumption information. By modeling the dynamic power of well-established components, we develop an integer linear programming (ILP)-based, encoding-agnostic, parametric power estimation model. Using this model, we demonstrate practical improvements in area and inter-crossbar communication by 0%–9.5% and 1.4%–5.1%, respectively. We also limit hotspot formation during optimization, achieving comparable or better results in targeted metrics with up to 96.4%–97.1% restriction of hotspot magnitude. Finally, we introduce profile-guided formulations to reduce worst-case and expected-case hotspot magnitude by 40.7%–69.5% and 40.6%–56.3%, respectively. Optimizing worst-case hotspot magnitude incidentally improves expected-case magnitude by 10.85%–33.45%. Reciprocally, optimizing expected-case magnitude incidentally improves worst-case magnitude by 4.33%–39.87%. Validation against hardware simulators confirms that PowerMappeR can decrease dynamic power consumption by 12.6%–27.3%.

Pohl, Devin [ORNL] (ORCID:0009000040149027)↗

An Augmented Lagrangian Filter Method for Real-Time Embedded Optimization

We present a filter line-search algorithm for nonconvex continuous optimization that combines an augmented Lagrangian function and a constraint violation metric to accept and reject steps. The approach is motivated by real-time optimization applications that need to be executed on embedded computing platforms with limited memory and processor speeds. The proposed method enables primal–dual regularization of the linear algebra system that in turn permits the use of solution strategies with lower computing overheads. We prove that the proposed algorithm is globally convergent and we demonstrate the developments using a nonconvex real-time optimization application for a building heating, ventilation, and air conditioning system. Our numerical tests are performed on a standard processor and on an embedded platform. Lastly, we demonstrate that the approach reduces solution times by a factor of over 1000.

97 MATHEMATICS AND COMPUTING↗

Magnetostatic focal spot correction for x-ray tubes operating in strong magnetic fields using iterative optimization

Purpose: Combining x-ray fluoroscopy and MR imaging systems for guidance of interventional procedures has become more commonplace. By designing an x-ray tube that is immune to the magnetic fields outside of the MR bore, the two systems can be placed in close proximity to each other. A major obstacle to robust x-ray tube design is correcting for the effects of the magnetic fields on the x-ray tube focal spot. A potential solution is to design active shielding that locally cancels the magnetic fields near the focal spot. Methods: An iterative optimization algorithm is implemented to design resistive active shielding coils that will be placed outside the x-ray tube insert. The optimization procedure attempts to minimize the power consumption of the shielding coils while satisfying magnetic field homogeneity constraints. The algorithm is composed of a linear programming step and a nonlinear programming step that are interleaved with each other. The coil results are verified using a finite element space charge simulation of the electron beam inside the x-ray tube. To alleviate heating concerns an optimized coil solution is derived that includes a neodymium permanent magnet. Any demagnetization of the permanent magnet is calculated prior to solving for the optimized coils. The temperature dynamics of the coil solutions are calculated using a lumped parameter model, which is used to estimate operation times of the coils before temperature failure. Results: For a magnetic field strength of 88 mT, the algorithm solves for coils that consume 588 A/cm{sup 2}. This specific coil geometry can operate for 15 min continuously before reaching temperature failure. By including a neodymium magnet in the design the current density drops to 337 A/cm{sup 2}, which increases the operation time to 59 min. Space charge simulations verify that the coil designs are effective, but for oblique x-ray tube geometries there is still distortion of the focal spot shape along with deflections of approximately 3 mm in the radial and circumferential directions on the anode. Conclusions: Active shielding is an attractive solution for correcting the effects of magnetic fields on the x-ray focal spot. If extremely long fluoroscopic exposure times are required, longer operation times can be achieved by including a permanent magnet with the active shielding design.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Bayesian optimization with active learning of design constraints using an entropy-based approach

Abstract The design of alloys for use in gas turbine engine blades is a complex task that involves balancing multiple objectives and constraints. Candidate alloys must be ductile at room temperature and retain their yield strength at high temperatures, as well as possess low density, high thermal conductivity, narrow solidification range, high solidus temperature, and a small linear thermal expansion coefficient. Traditional Integrated Computational Materials Engineering (ICME) methods are not sufficient for exploring combinatorially-vast alloy design spaces, optimizing for multiple objectives, nor ensuring that multiple constraints are met. In this work, we propose an approach for solving a constrained multi-objective materials design problem over a large composition space, specifically focusing on the Mo-Nb-Ti-V-W system as a representative Multi-Principal Element Alloy (MPEA) for potential use in next-generation gas turbine blades. Our approach is able to learn and adapt to unknown constraints in the design space, making decisions about the best course of action at each stage of the process. As a result, we identify 21 Pareto-optimal alloys that satisfy all constraints. Our proposed framework is significantly more efficient and faster than a brute force approach.

36 MATERIALS SCIENCE↗

First-principles study of structural stability, electronic, optical and elastic properties of binary intermetallic: PtZr

Structural, electronic, optical and elastic properties of PtZr have been studied using the full-potential linearized augmented plane wave (FP-LAPW) method within density functional theory (DFT). The energy against volume and enthalpy vs. pressure variation in three different structures i.e. B{sub 1}, B{sub 2} and B{sub 3} for PtZr has been presented. The equilibrium lattice parameter, bulk modulus and its pressure derivative have been obtained using optimization method for all the three phases. Furthermore, electronic structure was discussed to reveal the metallic character of the present compound. The linear optical properties are also studied under zero pressure for the first time. Results on elastic properties are obtained using generalized gradient approximation (GGA) for exchange correlation potentials. Ductile nature of PtZr compound is predicted in accordance with Pugh’s criteria.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Bi-fidelity Gradient-Based Approach for Nonlinear Well Logging Inverse Problems

Solving a non-linear inverse problem is challenging in computational science and engineering. Sampling based methods require a large number of model valuations; gradientbased methods require fewer model evaluations but only find the local minima. Multifidelity optimization combines the low fidelity model and the high-fidelity model to achieve both high accuracy and high efficiency. In this paper, we present a bi-fidelity approach to solve non-linear inverse problems. In the bi-fidelity inversion method, the low-fidelity model is used to acquire a good initial guess, and the high-fidelity model is used to locate the global minimum. Combined with a multi-start optimization scheme, the proposed approach significantly increases the possibility of finding the global minimum for nonlinear inverse problems with many local minima. The method is tested with two toy problems and then applied to an electromagnetic well logging inverse problem, which is difficult to solve using traditional gradient-based methods. The bi-fidelity method provides promising inversion results and can be easily applied to traditional gradient-based methods.

42 ENGINEERING↗

SU-F-T-01: Optimization of the Accelerated Partial Breast Brachytherapy Fractionation with Consideration of Physical Doses to Tumor and Organ at Risk

Purpose: The accelerated partial breast irradiation (APBI) with brachytherapy prescribes 34Gy to be delivered in 10 fractions over 5 consecutive working days without considering the physical dose to the target and organs at risk (OARs) for an individual patient. The purpose of this study is to optimize the fractionation scheme by evaluating the radiation effect on tumor and OARs with a modified linear-quadratic (LQ) model based on dose-volume histograms (DVHs). Methods: Five breast patients treated with multilumen balloon brachytherapy were selected. The minimum skin and rib spacing were ranged from 2.5mm to 14.3mm and from 1.0mm to 25.0mm, respectively. The LQ model parameters were set as: (1) breast: α=0.08, β=0.028, doubling time Tpot=14.4 days, and starting time Tk=21days; (2) skin: acute reaction α=0.101, β=0.009; late reaction α=0.064, β=0.029; (3) rib: α=0.3, β=0.12. Boundary dose Dt was 6 Gy for both target and OARs. The relation between radiation effects on the tumor (ET) and OARs (EOAR) were plotted for fraction number from 1 to 20. Results: The value of radiation effect from routine 3.4Gyx10 fractions was used as reference, ETref and EOARref. If set ET=ETref, the fractionation that results in minimum EOAR values correspond to the optimal fractionation. For these patients, the optimal numbers are 10 fractions for skin acute reaction, 18 fractions for skin and rib late reaction while the doses per fraction are 3.4Gy and 2.05–2.10Gy, respectively. If set EOAR=EOARref, the fractionation that results in a maximum ET value corresponds to the optimal fractionation. The optimal fractionation is 3.4Gyx10 fractions for skin acute reaction, and 2.10–2.25Gyx18 fractions for skin late reaction and rib. Conclusion: For APBI brachytherapy, the routine 3.4Gyx10 fractions is optimal fractionation for skin acute reaction, while 2.05–2.25Gyx18 fractions is optimal fractionation for late reaction of skin and rib.

60 APPLIED LIFE SCIENCES↗

RandONets: Shallow networks with random projections for learning linear and nonlinear operators

Deep neural networks have been extensively used for the solution of both the forward and the inverse problem for dynamical systems. However, their implementation necessitates optimizing a high-dimensional space of parameters and hyperparameters. This fact, along with the requirement of substantial computational resources, pose a barrier to achieving high numerical accuracy, but also interpretability. Here, to address the above challenges, we present Random Projection-based Operator Networks (RandONets): shallow networks with random projections and tailor-made numerical analysis methods that learn accurately and fast linear and nonlinear operators. Building on previous works, we prove that RandOnets are universal approximators of linear and nonlinear operators. Due to their simplicity, RandONets provide a one-step transformation of the input space, facilitating interpretability. For the evaluation of their performance, we focus on operators of PDEs. We show, that RandONets outperform by several orders of magnitude, both in terms of numerical approximation accuracy and computational cost, the “vanilla” DeepONets. Hence, we believe that our method will trigger further developments in the field of scientific machine learning, for the development of new ‘’light”schemes that will provide high accuracy while reducing dramatically the computational cost. A MATLAB toolbox for RandONets, including demos, is available on GitHub at https://github.com/GianlucaFabiani/RandONets.

Interpretable machine learning↗

Distributionally Robust Bilevel Optimization Model for Distribution Network With Demand Response Under Uncertain Renewables Using Wasserstein Metrics

Here, we consider a distribution network integrating demand response (DR) participants in the presence of uncertain renewable suppliers and outdoor temperatures. A bilevel optimization model is proposed to capture the intricate dynamics between price-incentivized DR participants and distribution system operations, including energy procurement and active/reactive power flows. The model is formulated as a distributional robust bilevel optimization using Wasserstein metrics. We show favorable data-driven properties including out-of-sample guarantee and asymptotic consistency. Furthermore, we present a tractable mixed-integer linear programming reformulation and characterize the worst-case distribution. Computational experiments are conducted on a modified 33-bus system. Our findings underscore the efficacy of the pricing strategies derived from the proposed bilevel optimization model. These strategies not only effectively manage DR participants' behavior but also bring equity considerations among households with various characteristics to light. The results contribute to a deeper understanding of the interplay between distribution system operators and DR participants.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Optimization of Dynamic Aperture of PEP-X Baseline Design

SLAC is developing a long-range plan to transfer the evolving scientific programs at SSRL from the SPEAR3 light source to a much higher performing photon source. Storage ring design is one of the possibilities that would be housed in the 2.2-km PEP-II tunnel. The design goal of PEPX storage ring is to approach an optimal light source design with horizontal emittance less than 100 pm and vertical emittance of 8 pm to reach the diffraction limit of 1-{angstrom} x-ray. The low emittance design requires a lattice with strong focusing leading to high natural chromaticity and therefore to strong sextupoles. The latter caused reduction of dynamic aperture. The dynamic aperture requirement for horizontal injection at injection point is about 10 mm. In order to achieve the desired dynamic aperture the transverse non-linearity of PEP-X is studied. The program LEGO is used to simulate the particle motion. The technique of frequency map is used to analyze the nonlinear behavior. The effect of the non-linearity is tried to minimize at the given constrains of limited space. The details and results of dynamic aperture optimization are discussed in this paper.

43 PARTICLE ACCELERATORS↗

Calculation of rf current drive in tokamaks

The toroidal plasma current is a key parameter for controlling MHD stability and fusion performances in tokamaks. Among the various methods for driving current, rf waves are a flexible and powerful tool. Therefore, their role in the design and optimization of advanced scenarios is considerable. The universal ray-tracing code C3PO coupled with the fully implicit linearized 3-D bounce-averaged relativistic electron Fokker-Planck solver LUKE is an illustration of the present day effort for performing fast and realistic calculations of the rf driven plasma current. The versatility of this tool is highlighted by simulations concerning the lower hybrid and electron cyclotron waves.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Design of Distributed Controllers Seeking Optimal Power Flow Solutions Under Communication Constraints

This paper focuses on power distribution networks featuring inverter-interfaced distributed energy resources (DERs), and develops feedback controllers that drive the DER output powers to solutions of time-varying AC optimal power flow (OPF) problems. Control synthesis is grounded on primal-dual-type methods for regularized Lagrangian functions, as well as linear approximations of the AC power-flow equations. Convergence and OPF-solution-tracking capabilities are established while acknowledging: i) communication-packet losses, and ii) partial updates of control signals. The latter case is particularly relevant since it enables asynchronous operation of the controllers where DER setpoints are updated at a fast time scale based on local voltage measurements, and information on the network state is utilized if and when available, based on communication constraints. As an application, the paper considers distribution systems with high photovoltaic integration, and demonstrates that the proposed framework provides fast voltage-regulation capabilities, while enabling the near real-time pursuit of solutions of AC OPF problems.

24 POWER TRANSMISSION AND DISTRIBUTION↗