Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “uncertainty”

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 829 records · Page 46

Inverse Uncertainty Quantification of Reactor Simulation with Polynomial Chaos Surrogate Model

Reliable uncertainty quantification (UQ) and sensitivity analysis (SA) is significant part of modelling and simulation. This is particularly important in the nuclear community due to the fact that the licensing of nuclear installations has shifted from conservative to Best Estimate plus Uncertainty (BEPU) methodologies. UQ can be broadly classified as forward UQ and inverse UQ. Forward UQ is the process of quantifying the uncertainties in simulation outputs by propagating the uncertainties in input parameters. It requires knowledge in the input uncertainty, which is often specified by ad-hoc expert judgment which lacks mathematical rigor. The inverse UQ is the process to determine the uncertainty in input parameters given observations of the output. The purpose of this paper is to replace such ad-hoc expert judgment of the statistical properties of input model parameters in nuclear reactor simulation. This summary presents work on inverse UQ of a nonlinear Point Reactor Kinetics Equations - lumped thermal-hydraulics parameter coupled problem. Bayesian inference and Markov Chain Monte Carlo (MCMC) are used for the inverse UQ. Because MCMC requires a large number of simulations which is impractical for models that are computationally expensive, surrogate models is constructed by generalized Polynomial Chaos (PC) expansion technique. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Uncertainty Quantification with MCMC method and BFBT Data

The reliability of predictions of the system codes is closely related to the validation of their physical models. For example, the accuracy of void fraction prediction in a Boiling Water Reactor (BWR) using TRACE code depends on the uncertainties of closure relations developed for the two-phase flow models. This work focus on quantifying the uncertainties of two physical models, interfacial drag model and subcooled boiling model, using Markov Chain Monte Carlo (MCMC) method and BFBT experimental benchmark data. We applied the MCMC and MLE algorithms to BFBT benchmark data and estimated the uncertainty distribution of two model parameters: subcooled boiling heat transfer coefficient and interfacial drag (bubbly/slug rod bundle) coefficient. These algorithms tend to give various estimation results. Despite the differences in these algorithms, we find that the estimation results are consistent. After obtaining the uncertainty estimates, we validated the uncertainty estimates by applying the estimated distributions of model parameters to TRACE prediction. The new predictions succeed in reducing the void fraction prediction error, but it is not possible to eliminate all prediction error. We are only adjusting two model parameters, and there exists uncertainties in all model parameters. Application of the algorithms to other physical models and experimental data will be helpful for future work. (authors)

42 ENGINEERING↗

Calibration and Uncertainty Estimation Using the Ensemble Kalman Filter with a Large Subsurface Flow and Transport Model - 20321

At routinely monitored groundwater contamination sites, periodically measured environmental conditions such as groundwater levels and contaminant concentrations are used to inform and confirm a conceptual site model (CSM) and guide the development and calibration of a numerical groundwater flow and transport model. The calibration of groundwater flow and transport models after each measurement (sampling) event can illuminate deficiencies in a CSM, identify areas where additional monitoring is warranted, and predict the behavior of the system to guide decision making. However, manual and automated (e.g. PEST) model calibration tools can be time-consuming and computationally expensive to implement after each sampling event. Perhaps as a result, such calibration tools generally utilize all available monitoring data simultaneously rather than sequentially assimilating monitoring data one sampling event at a time as the results from sampling become available. A more real-time data assimilation approach may reduce parameter uncertainty, quantify the value of additional monitoring data, and produce a usable model more quickly and with less effort. To mitigate the potential time-consuming aspects of manual and widely applied automated calibration techniques, a data assimilation algorithm called the ensemble Kalman filter (EnKF) was evaluated as a relatively efficient method of model calibration and uncertainty assessment via the sequential integration of monitoring data into a model. The EnKF was able to successfully and efficiently assimilate monitoring and modeling data to calibrate a complex flow and transport model at a real-world site with significant subsurface heterogeneity, uncertainty, and 12 years of monitoring data (over 4,000 individual measurements of groundwater levels and over 2,500 measurements of contaminant concentrations). Starting with an uncalibrated model data from annual sampling events were sequentially assimilated, and the resultant predication errors and estimated parameter uncertainties were tracked. After all monitoring data were assimilated, both flow and transport residuals at the end of the EnKF process were comparable to those produced via a concurrent PEST calibration effort but required fewer model simulations. Both uncertainty and prediction errors decreased over time. In a real-time application, the adequacy of the model could be assessed after each sampling event. The benefits of such a real-time approach to utilizing monitoring data include reduced costs (in the form of model updates or site characterization efforts), early flagging of possible errors in the CSM, and a reduced risk of overfitting and corresponding increased confidence in model predictions. This tool may be particularly useful compared to other calibration techniques (e.g. manual, PEST) when model runtimes are long, calibration parameters are many, or parameter uncertainty is large. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

use of cross-section covariance information in uncertainty analysis at modeling HTGR neutron physics - 18551

This work is aimed at improvement of high temperature gas cooled reactors (HTGR) neutron physics design calculations by application of uncertainty analysis with the use of cross-section covariance information. Methodology and codes for preparation of multigroup libraries of covariance information for individual isotopes from the basic 44-group binary library of SCALE-6 code system were developed. A 69-group library of covariance information in a special format for main isotopes and elements typical for HTGR was generated. This library can be used for estimation of uncertainties associated with nuclear data in analysis of HTGR neutron physics with design codes. As an example, calculations of one-group cross-section uncertainties for fission and capture reactions for main isotopes of the MHTGR-350 benchmark, as well as uncertainties of the multiplication factor (k{sub ∞}) for the MHTGR-350 fuel compact cell model and fuel block model were performed. These uncertainties were estimated by the developed technology with the use of WIMS-D code (technology SUW) and modules of SCALE-6 code system, namely, by TSUNAMI, KENO-VI and SAMS. The eight most important reactions on isotopes for MHTGR-350 benchmark were identified, namely: {sup 10}B(capt), {sup 238}U(n,γ), ν{sub 5}, {sup 235}U(n,γ), {sup 238}U(el), {sup nat}C(el), {sup 235}U(fiss)-{sup 235}U(n,γ) and {sup 235}U(fiss). It is shown that the main contribution in the uncertainty of multiplication factor of the fuel block model gives the reaction {sup 10}B(capt) ((n,α)+(n,γ)). (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Quantifying Transient Uncertainty in the BEAVRS Benchmark Using Time Series Analysis Methods

Advances in computation have brought about significant improvements in creating fast-running high-fidelity simulations of nuclear cores. The BEAVRS benchmark is a highly-detailed PWR specification with two cycles of measured operational data used to validate high-fidelity core analysis methods. This PWR depletion benchmark captures the fine details of the LWR fuel assemblies, burnable absorbers, in-core fission detectors, core loading and shuffling patterns. Specifically, 58 of the 193 assemblies contain in-core detectors with measurements taken over 61 axial positions every month. These detectors are U-235 fission chambers with slightly varying mass of U-235. The collected signals are normalized on a given assembly permitting full core comparisons. The fuel layout for cycle 1 and instrument tube locations for the reactor are given in figures 1 and 2 respectively. Through a series of data processing and comparisons, it was shown that axially integrated radial maps of reaction rates were in close agreement between provided detector data and calculated data. More recently, the focus of the BEAVRS project has been on quantifying uncertainty to further assess the validity of aforementioned results. A close investigation of sources of error shows that uncertainties from operational nuclear data arise primarily from data measurements and processing. While the errors in axial realignment and interpolation seem significant, these sources affect only a small subset of data points near grid spacers and do not persist throughout the entire core. Instead, measurement uncertainty from detector count rates and radial integration dominate the uncertainty in measured data are determined as the most significant sources of error. This uncertainty analysis treated each given burnup step as independent of comparable data at neighboring burnup steps. More recent work has been targeted at understanding how reaction rates vary over time, in order to determine whether calculated reaction rates follow any observable trends. This abstract hones in on transient uncertainty quantification, in order to compare observed data against models for transient behavior. The first section of the paper uses linear regression tools to fit operational data, while the latter portion of the paper explores more complex simulation tools to fit operational data. Ultimately, the BEAVRS benchmark aims to serve as a true non-proprietary international benchmark for the validation of high-fidelity tools. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Uncertainty due to Nuclear Data for an MTR Fuel Assembly

For a research reactor, and in particular for the High Flux Reactor (HFR) in Petten, every cycle has a distinct core loading. This regards both the fuel assemblies loading and mainly the loading of the experiment rigs. These experiments can be of the most diverse types, comprising strong absorbing materials as well as fissile isotopes. This poses some challenges at every cycle because of the possible influence on the power profile and neutron flux distribution, which are of great importance for the isotopic production facilities loaded at every cycle in the case of the HFR. The isotope production facilities require predictable and preferably as constant as possible irradiation conditions (neutron flux levels, mainly). Another point of interest is the cycle length, which has to be guaranteed beforehand by taking into account in the assessment analysis the expected on-power core loading changes. These changes are caused by regular loading/unloading of some of the isotope production rigs. For optimization reasons, there is also a tendency to increase the cycle lengths and at the same time to minimize the costs associated to the loading of fresh fuel and control rod assemblies. To justify a further increase in cycle length, a quantification of the uncertainties on the reactivity curve is required, so that the adopted design margins can be decreased in a controlled and responsible manner. The uncertainties in the reactivity curve can be associated to several sources, ranging from uncertainties in engineering and operational parameters to uncertainties in basic nuclear data. This paper includes the quantification of uncertainty on the reactivity curve associated to a typical HFR fuel assembly as result of uncertainties in the cross section data of the main fuel, structural and coolant isotopes. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Uncertainty Quantification by Monte Carlo Analysis using CFD Simulations for GEMIX Benchmark Activities

Uncertainty quantification (UQ) methods can be applied to computational fluid dynamics (CFD) simulations for determining the uncertainty bands of the simulations results. In order to determine the applicability of these different methods, simple benchmark activities have been conducted by various group for comparison. The GEMIX (Generic Mixing experiment) test facility was used to develop benchmark data for this purpose by the Paul Scherrer Institute (PSI). The facility involves the mixing of two fluids that are initially separated. One of the two fluids is water with sugar dissolved in it to create a fluid with similar fluid properties to water, except for density. The other fluid was distilled water to reduce the uncertainty in the fluid's properties. During the experiments, velocity fields and concentration measurements were acquired at several planes of interest. The velocity fields were acquired using particle image velocimetry (PIV) and concentrations of each fluid were acquired using laser induced fluorescence (LIF). The inlet velocity profiles were acquired using laser Doppler velocimetry (LDV). The activity provided the opportunity to vary turbulence models, meshing methodologies, and numerical schemes for creation of uncertainty estimates. The present study investigates the usage of a simple Monte Carlo (MC) technique for varying the inlet boundary conditions of each inlet. The velocity profiles and densities of each inlet were specifically varied based on experiment measurement uncertainty. The intent was to isolate the uncertainty due to inlet quantities and not related to meshing, numerical scheme, or turbulence models. The comparisons of experimental and simulation results were found to be reasonably accurate. Though deficiencies were identified in the modeling approach to be corrected in future works by the authors. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Uncertainty and Sensitivity Analysis of TREAT Transient Test no. 15

The U.S. Department of Energy (DOE) is preparing for the resumption of transient testing using the Transient Reactor Test Facility (TREAT) located at the Idaho National Laboratory (INL). In order to support the analysis of TREAT operation and experiment planning, the MAMMOTH reactor physics team has recently begun development of uncertainty and sensitivity analysis for the TREAT core simulated via MAMMOTH. MAMMOTH is a high-fidelity three-dimensional transient multi-physics code developed at INL under the DOE-NE Advanced Modeling and Simulation program (NEAMS). This research is part of the development presenting some initial uncertainty and sensitivity analysis of TREAT Transient Test no. 15. In this summary, we employ the Sobol sampling strategy implemented in RAVEN to efficiently perform the sensitivity and uncertainty analysis. The idea is using Sobol functional decomposition with generalized polynomial chaos (gPC) expansions to estimate the responses of interest. The gPC coefficients are then computed via the sparse grid spectral projection method. In addition, this strategy includes evaluation of the second or higher order sensitivity indices, allowing the study of interactions between several input variables. The principal focus of this analysis is to quantify the uncertainties in the total power and maximum fuel temperatures with respect to uncertainties in the initial transient conditions, such as rated power, fuel temperature, graphite fuel density and control rod boron concentration. In addition, both the first order and second order time-dependent sensitivity indices are computed and analyzed. The results show that the interactions between initial control rod boron concentration and the rated power will contribute at most 1% to the total uncertainty of the power peak.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Towards application of a neutron cross-section uncertainty propagation capability in the criticality safety methodology - 14434

At the Paul Scherrer Institute (PSI) the methodology for criticality safety evaluations (CSE) used for Swiss applications is focused, following common practice, on establishing an upper subcriticality limit (USL) for the effective neutron multiplication factor (k{sub eff}), based on a statistical evaluation of the obtained validation results in terms of the calculated-to-benchmark k{sub eff} values. Recently, a new computational tool referred to as NUSS (Nuclear data Uncertainty Stochastic Sampling) was developed in order to propagate neutron-nuclear data (ND) uncertainties in calculations with Monte Carlo transport codes in combination with general-purpose ND libraries in ACE format. Thereby, this tool should now allow for a quantification of ND data uncertainties in CSE and to introduce these into the CSE criteria definition. This paper concerns an application of the NUSS methodology for criticality validation studies. Calculated k{sub eff} uncertainty estimations were obtained for the current PSI validation suite including in total 149 critical benchmark configurations for which in-house MCNP(X) models were developed, covering both uranium and mixed uranium-plutonium oxide fuels. The MCNPX-2.7.0 code was used together with the ENDF/B-VII.1 ND library while the ND variance-covariance matrixes (VCM) were taken from both SCALE-6.0 44-group libraries and also from ENDF/B-VII.1 files. Based on the newly obtained information on the uncertainties, a re-assessment of the lower tolerance bound (LTB) for k{sub eff} using the total uncertainties as weighting factors is proposed as means for the present PSI CSE methodology upgrade and some possible directions of further methodology enhancements are discussed next. (authors)

42 ENGINEERING↗

Impact of fission yield covariance matrices on decay heat uncertainty quantification with the DARWIN2 package

Although fission yields are strongly correlated, correlations between them are usually not taken into account when performing decay heat uncertainty calculations with the CEA DARWIN2 package due to the lack of reference covariance matrices associated to the JEFF-3.1.1 evaluation. However, covariance matrices for {sup 235}U and {sup 239}Pu thermal fission yields have been produced recently by the subgroup 37 of OECD/NEA Working Party on International Nuclear Data Evaluation Cooperation (WPEC). The study presented in this paper evaluates the effect of those covariances on the decay heat uncertainty calculation, both on fission burst experiments and on integral decay heat measurements that are part of the DARWIN2 experimental validation database. Although some differences are observed, partly due to the variety of models used to produce the matrices, the propagation of fission yield covariances always leads to a reduction of the decay heat uncertainty for cooling time above 10 seconds, indicating that not taking them into account is a conservative hypothesis from a safety point of view. Nevertheless, the strong impact of those covariances on the decay heat uncertainty also highlights the need of documented and consistent fission yield uncertainties and correlation data. On top of that, all the covariance matrices used for this study represent the correlation between the physical model parameters, but none of them includes the experimental correlations. An evaluation of the experimental correlations would also be of strong interest for decay heat uncertainty calculations. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Data-Driven Computation of Probabilistic Marching Cubes for Efficient Visualization of Level-Set Uncertainty

Uncertainty visualization is an important emerging research area. Being able to visualize data uncertainty can help scientists improve trust in analysis and decision-making. However, visualizing uncertainty can add computational overhead, which can hinder the efficiency of analysis. In this paper, we propose novel data-driven techniques to reduce the computational requirements of the probabilistic marching cubes (PMC) algorithm. PMC is an uncertainty visualization technique that studies how uncertainty in data affects level-set positions. However, the algorithm relies on expensive Monte Carlo (MC) sampling for the multivariate Gaussian uncertainty model because no closed-form solution exists for the integration of multivariate Gaussian. In this work, we propose the eigenvalue decomposition and adaptive probability model techniques that reduce the amount of MC sampling in the original PMC algorithm and hence speed up the computations. Our proposed methods produce results that show negligible differences compared with the original PMC algorithm demonstrated through metrics, including root mean squared error, maximum error, and difference images. We demonstrate the performance and accuracy evaluations of our data-driven methods through experiments on synthetic and real datasets.

Athawale, Tushar↗

A comparison of approximate reasoning results using information uncertainty

An Approximate Reasoning (AR) model is a useful alternative to a probabilistic model when there is a need to draw conclusions from information that is qualitative. For certain systems, much of the information available is elicited from subject matter experts (SME). One such example is the risk of attack on a particular facility by a pernicious adversary. In this example there are several avenues of attack, i.e. scenarios, and AR can be used to model the risk of attack associated with each scenario. The qualitative information available and provided by the SME is comprised of linguistic values which are well suited for an AR model but meager for other modeling approaches. AR models can produce many competing results. Associated with each competing AR result is a vector of linguistic values and a respective degree of membership in each value. A suitable means to compare and segregate AR results would be an invaluable tool to analysts and decisions makers. A viable method would be to quantify the information uncertainty present in each AR result then use the measured quantity comparatively. One issue of concern for measuring the infornlation uncertainty involved with fuzzy uncertainty is that previously proposed approaches focus on the information uncertainty involved within the entire fuzzy set. This paper proposes extending measures of information uncertainty to AR results, which involve only one degree of membership for each fuzzy set included in the AR result. An approach to quantify the information uncertainty in the AR result is presented.

45 MILITARY TECHNOLOGY, WEAPONRY, AND NATIONAL DEF↗

Fuzzy-probabilistic calculations of water-balance uncertainty

Hydrogeological systems are often characterized by imprecise, vague, inconsistent, incomplete, or subjective information, which may limit the application of conventional stochastic methods in predicting hydrogeologic conditions and associated uncertainty. Instead, redictions and uncertainty analysis can be made using uncertain input parameters expressed as probability boxes, intervals, and fuzzy numbers. The objective of this paper is to present the theory for, and a case study as an application of, the fuzzyprobabilistic approach, ombining probability and possibility theory for simulating soil water balance and assessing associated uncertainty in the components of a simple waterbalance equation. The application of this approach is demonstrated using calculations with the RAMAS Risk Calc code, to ssess the propagation of uncertainty in calculating potential evapotranspiration, actual evapotranspiration, and infiltration-in a case study at the Hanford site, Washington, USA. Propagation of uncertainty into the results of water-balance calculations was evaluated by hanging he types of models of uncertainty incorporated into various input parameters. The results of these fuzzy-probabilistic calculations are compared to the conventional Monte Carlo simulation approach and estimates from field observations at the Hanford site.

54 ENVIRONMENTAL SCIENCES↗

Modern Monte Carlo methods for efficient uncertainty quantification and propagation: A survey

Uncertainty quantification (UQ) includes the characterization, integration, and propagation of uncertainties that result from stochastic variations and a lack of knowledge or data in the natural world. Monte Carlo (MC) method is a sampling-based approach that has widely used for quantification and propagation of uncertainties. However, the standard MC method is often time-consuming if the simulation-based model is computationally intensive. This article gives an overview of modern MC methods to address the existing challenges of the standard MC in the context of UQ. Specifically, multilevel Monte Carlo (MLMC) extending the concept of control variates achieves a significant reduction of the computational cost by performing most evaluations with low accuracy and corresponding low cost, and relatively few evaluations at high accuracy and corresponding high cost. Multifidelity Monte Carlo (MFMC) accelerates the convergence of standard Monte Carlo by generalizing the control variates with different models having varying fidelities and varying computational costs. Multimodel Monte Carlo method (MMMC), having a different setting of MLMC and MFMC, aims to address the issue of UQ and propagation when data for characterizing probability distributions are limited. Multimodel inference combined with importance sampling is proposed for quantifying and efficiently propagating the uncertainties resulting from small data sets. All of these three modern MC methods achieve a significant improvement of computational efficiency for probabilistic UQ, particularly uncertainty propagation. Finally, an algorithm summary and the corresponding code implementation are provided for each of the modern MC methods. The extension and application of these methods are discussed in detail.

97 MATHEMATICS AND COMPUTING↗

Stochastic scheduling for commercial building cooling systems: considering uncertainty in zone temperature prediction

Here, this paper presents the first attempt to address the uncertainty in zone temperature prediction with stochastic optimization. The uncertain zone temperature is a process uncertainty and has not been considered in the existing stochastic optimization for building control. To fill this gap, we proposed a novel formulation of stochastic optimization to handle process uncertainty in building control. Specifically, we first examined the accuracy of a typical linear model for predicting zone temperature. We then formulated the scheduling of the building cooling system as a stochastic optimization problem over a 24-hour look-ahead period to minimize the electricity cost of the studied building cooling system. After that, we applied the proposed stochastic load scheduling (SLS) to a direct expansion (DX) cooling system that serves a medium office building. Through simulation with a detailed building energy simulation software, EnergyPlus, we evaluated the operational cost and the thermal comfort compared with a deterministic load scheduling. The operation cost of scheduling was found to vary with the level of zone temperature prediction uncertainty. The proposed SLS can mitigate the impacts of uncertain zone temperature predictions on both operational cost and thermal comfort. The evaluation results indicate that the proposed SLS works better when the uncertainty level is more significant.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Uncertainty quantification and propagation in CALPHAD modeling

Design is about making decisions bounded by a quantifiable degree of certainty. In the context of alloy design, Integrated Computational Materials Engineering (ICME) provides the framework whereby performance requirements are ultimately transformed into alloy/processing specifications through the combination of (complex) computational models connecting process-structure-property-performance relationships and experiments. Most ICME approaches consider the models used as deterministic and thus do not provide the means to make alloy design decisions with proper confidence measures. At the root of ICME lie CALPHAD models that describe the thermodynamics and phase stability of phases under specific thermodynamic boundary conditions. To date, the vast majority of efforts within the CALPHAD community have been deterministic in that thermodynamic models and the resulting thermodynamic properties and phase diagram features do not explicitly account for the uncertainties inherent in the model formulation or in the experimental/computational data used. In this contribution, we provide an overview of the state of the field. We review major efforts thus far and we then provide a (brief) tutorial on basic concepts of uncertainty quantification and propagation (UQ/UP) in CALPHAD. We discuss the major features of frequentist and Bayesian interpretations of uncertainty and proceed with a discussion of recent case studies in which UQ has been used to parameterize models for the thermodynamic properties of phases. We follow our discussion by presenting frameworks and demonstrating the propagation of uncertainty in thermodynamic properties and phase diagram predictions and briefly discuss how we can use Bayesian frameworks for rigorous model selection as well as for model fusion. Furthermore, we close our contribution by providing context for what has been done and what remains to be accomplished in order to fully embrace the management of uncertainty in CALPHAD modeling, a foundational element of ICME.

42 ENGINEERING↗

Optimal Wind Power Uncertainty Intervals for Electricity Market Operation

It is important to select an appropriate uncertainty level of the wind power forecast for power system scheduling and electricity market operation. Traditional methods hedge against a predefined level of wind power uncertainty, such as a specific confidence interval or uncertainty set, which leaves the questions of how to best select the appropriate uncertainty levels. To bridge this gap, this paper proposes a model to optimize the forecast uncertainty intervals of wind power for power system scheduling problems, with the aim of achieving the best trade-off between economics and reliability. Then we reformulate and linearize the models into a mixed integer linear programming (MILP) without strong assumptions on the shape of the probability distribution. In order to invest the impacts on cost, reliability, and prices in a electricity market, we apply the proposed model on a twosettlement electricity market based on a six-bus test system and on a power system representing the U.S. state of Illinois. The results show that the proposed method can not only help to balance the economics and reliability of the power system scheduling, but also help to stabilize the energy prices in electricity market operation.

electricity market↗

Multilevel Monte Carlo Estimators For Derivative-Free Optimization Under Uncertainty

Optimization is a key tool for scientific and engineering applications; however, in the presence of models affected by uncertainty, the optimization formulation needs to be extended to consider statistics of the quantity of interest. Optimization under uncertainty (OUU) deals with this endeavor and requires uncertainty quantification analyses at several design locations; i.e., its overall computational cost is proportional to the cost of performing a forward uncertainty analysis at each design location. An OUU workflow has two main components: an inner loop strategy for the computation of statistics of the quantity of interest, and an outer loop optimization strategy tasked with finding the optimal design, given a merit function based on the inner loop statistics. Here, in this work, we propose to alleviate the cost of the inner loop uncertainty analysis by leveraging the so-called multilevel Monte Carlo (MLMC) method, which is able to allocate resources over multiple models with varying accuracy and cost. The resource allocation problem in MLMC is formulated by minimizing the computational cost given a target variance for the estimator. We consider MLMC estimators for statistics usually employed in OUU workflows and solve the corresponding allocation problem. For the outer loop, we consider a derivative-free optimization strategy implemented in the SNOWPAC library; our novel strategy is implemented and released in the Dakota software toolkit. We discuss several numerical test cases to showcase the features and performance of our approach with respect to its Monte Carlo single fidelity counterpart.

97 MATHEMATICS AND COMPUTING↗