Search NASA⌕ Search

SEARCH · Search NASA

Results for “Stochastic System”

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 217 records · Page 12

Operation Optimization Using Reinforcement Learning with Integrated Artificial Reasoning Framework

In large and complex systems, operational decision-making requires a systematic analysis with a vast amount of data from both process parameters and component status monitoring. In this paper, we present an integrated artificial reasoning approach for system state transition models that can help operational decision-making with explainable and traceable reasoning. The integrated artificial reasoning framework is a physics-based approach of defining the system structure in a Bayesian network, so we leveraged it in a Markov decision process (MDP) for finding optimal operational solutions. In our proposed framework, the MDP is implemented on a dynamic Bayesian network (DBN), which represents causalities in a system. The multilevel flow modeling was utilized in order to extract these causalities in a more efficient and objective manner. Since multilevel flow modeling is based on the fundamental energy and mass conservation laws, the target system is decomposed into several mass, energy, and information structures, which serve as the basis for a DBN. The MDP consists of the processes of finding a solution for the Bellman equation, which can be derived from the conditional probability equations of the constructed DBN. System operators can capture stochastic system dynamics as multiple subsystem state transitions based on their physical relations and uncertainties coming from the component degradation process or random failures. We analyzed a simplified example system to illustrate finding an optimal operational policy with this approach.

Kim, Junyung↗

Getting Astrophysical Information from LISA Data

Gravitational wave signals from a large number of astrophysical sources will be present in the LISA data. Information about as many sources as possible must be estimated from time series of strain measurements. Several types of signals are expected to be present: simple periodic signals from relatively stable binary systems, chirped signals from coalescing binary systems, complex waveforms from highly relativistic binary systems, stochastic backgrounds from galactic and extragalactic binary systems and possibly stochastic backgrounds from the early Universe. The orbital motion of the LISA antenna will modulate the phase and amplitude of all these signals, except the isotropic backgrounds and thereby give information on the directions of sources. Here we describe a candidate process for disentangling the gravitational wave signals and estimating the relevant astrophysical parameters from one year of LISA data. Nearly all of the sources will be identified by searching with templates based on source parameters and directions.

Stebbins, R. T.↗

Impact of Distributed Energy Resources on the Bulk Electric System: Combined Modeling of Transmission and Distribution Systems and Benchmark Case Studies

The high penetration of distributed energy resources (DERs) on modern distribution systems introduces intermittent power generation, stochastic system operating conditions, and bi-directional power flows that impact the system responses to various types of disturbances and may pose challenges to the operations and reliability of the Bulk Electric System (BES). In order to understand the impact of DERs on the planning and operation of the BES, the distribution system can no longer be represented as a single load at a node on the transmission system. At the same time, advanced control capabilities of DERs offer potential opportunities for improving BES reliability by transforming DERs from a passive “do no harm” resource to an active “support reliability” resource when applied in a planned and well-thought-out manner. This concept is often described as DER “flexibility.” With the persistent trend towards a high- penetration of DERs in distribution systems, regulators such as the North American Electric Reliability Corporation (NERC), planning coordinators and grid operators are in serious need of understanding the full-spectrum impact of DERs on the BES to ensure secure and reliable grid planning and operations.

24 POWER TRANSMISSION AND DISTRIBUTION↗

Pulsed acceleration in solar flares

We study the nonlinear dynamics of particle acceleration in solar flares by analyzing the time series of various quasi-periodic radio signatures during flares. In particular we present the radio and hard X-ray data of three flares which suppport the following tentative conclusions: (1) Particle acceleration and injection into magnetic structures occurs intrinsically in a pulsed mode (with a typical period of 1-2 s), produced by a single, spatially coherent, nonlinear system, rather than by a stochastic system with many spatially independent components ('statistical flare' produced by a fragmented primary energy release). (2) The nonlinear (quasi-periodic) mode of pulsed particle acceleration and injection into a coronal loop can be stabilized by phase locking with an MHD wave (oscillation) mode, if both periods are close to each other. (3) Pulsed injection of electron beams into a coronal loop may trigger nonlinear relaxational oscillations of wave-particle interactions. This is particularly likely when the limit cycles of both systems are similar.

Aschwanden, Markus J.↗

Dynamic Event Tree Analysis Through RAVEN

Conventional Event-Tree (ET) based methodologies are extensively used as tools to perform reliability and safety assessment of complex and critical engineering systems. One of the disadvantages of these methods is that timing/sequencing of events and system dynamics is not explicitly accounted for in the analysis. In order to overcome these limitations several techniques, also know as Dynamic Probabilistic Risk Assessment (D-PRA), have been developed. Monte-Carlo (MC) and Dynamic Event Tree (DET) are two of the most widely used D-PRA methodologies to perform safety assessment of Nuclear Power Plants (NPP). In the past two years, the Idaho National Laboratory (INL) has developed its own tool to perform Dynamic PRA: RAVEN (Reactor Analysis and Virtual control ENvironment). RAVEN has been designed in a high modular and pluggable way in order to enable easy integration of different programming languages (i.e., C++, Python) and coupling with other application including the ones based on the MOOSE framework, developed by INL as well. RAVEN performs two main tasks: 1) control logic driver for the new Thermo-Hydraulic code RELAP-7 and 2) post-processing tool. In the first task, RAVEN acts as a deterministic controller in which the set of control logic laws (user defined) monitors the RELAP-7 simulation and controls the activation of specific systems. Moreover, RAVEN also models stochastic events, such as components failures, and performs uncertainty quantification. Such stochastic modeling is employed by using both MC and DET algorithms. In the second task, RAVEN processes the large amount of data generated by RELAP-7 using data-mining based algorithms. This paper focuses on the first task and shows how it is possible to perform the analysis of dynamic stochastic systems using the newly developed RAVEN DET capability. As an example, the Dynamic PRA analysis, using Dynamic Event Tree, of a simplified pressurized water reactor for a Station Black-Out scenario is presented.

42 ENGINEERING↗

SDYN-GANs: Adversarial learning methods for multistep generative models for general order stochastic dynamics

We introduce adversarial learning methods for data-driven generative modeling of dynamics of nth-order stochastic systems. Our approach builds on Generative Adversarial Networks (GANs) with generative model classes based on stable m-step stochastic numerical integrators. From observations of trajectory samples, we introduce methods for learning long-time predictors and stable representations of the dynamics. Our approaches use discriminators based on Maximum Mean Discrepancy (MMD), training protocols using both conditional and marginal distributions, and methods for learning dynamic responses over different time-scales. We show how our approaches can be used for modeling physical systems to learn force-laws, damping coefficients, and noise-related parameters. Our adversarial learning approaches provide methods for obtaining stable generative models for dynamic tasks including long-time prediction and developing simulations for stochastic systems.

• Artificial intelligence (AI) / machine learning ↗

On the equivalence of dynamically orthogonal and bi-orthogonal methods: Theory and numerical simulations

The Karhunen–Lòeve (KL) decomposition provides a low-dimensional representation for random fields as it is optimal in the mean square sense. Although for many stochastic systems of practical interest, described by stochastic partial differential equations (SPDEs), solutions possess this low-dimensional character, they also have a strongly time-dependent form and to this end a fixed-in-time basis may not describe the solution in an efficient way. Motivated by this limitation of standard KL expansion, Sapsis and Lermusiaux (2009) [26] developed the dynamically orthogonal (DO) field equations which allow for the simultaneous evolution of both the spatial basis where uncertainty ‘lives’ but also the stochastic characteristics of uncertainty. Recently, Cheng et al. (2013) [28] introduced an alternative approach, the bi-orthogonal (BO) method, which performs the exact same tasks, i.e. it evolves the spatial basis and the stochastic characteristics of uncertainty. In the current work we examine the relation of the two approaches and we prove theoretically and illustrate numerically their equivalence, in the sense that one method is an exact reformulation of the other. We show this by deriving a linear and invertible transformation matrix described by a matrix differential equation that connects the BO and the DO solutions. We also examine a pathology of the BO equations that occurs when two eigenvalues of the solution cross, resulting in an instantaneous, infinite-speed, internal rotation of the computed spatial basis. We demonstrate that despite the instantaneous duration of the singularity this has important implications on the numerical performance of the BO approach. On the other hand, it is observed that the BO is more stable in nonlinear problems involving a relatively large number of modes. Several examples, linear and nonlinear, are presented to illustrate the DO and BO methods as well as their equivalence.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Numerical analysis of randomly forced glycolitic oscillations

Randomly forced glycolytic oscillations in Higgins model are studied both numerically and analytically. Numerical analysis is based on the direct simulation of the solutions of stochastic system. Non-uniformity of the stochastic bundle along the deterministic cycle is shown. For the analytical investigation of the randomly forced Higgins model, the stochastic sensitivity function technique and confidence domains method are applied. Results of the influence of additive noise on the cycle of this model are given.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Stochastic dynamics of a delayed bistable system with multiplicative noise

In this paper we investigate the properties of a delayed bistable system under the effect of multiplicative noise. We first prove the existence and uniqueness of the positive solution and show that its moments are uniformly bounded. Then, we study stochastic dynamics of the solution in long time, the lower and upper bounds for the paths and an estimate for the average value are provided.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Stochastic Energy Deployment System (SEDS)

This short report introduces the Stochastic Energy Deployment System (SEDS), an economy-wide energy model of the United States. Sections cover SEDS methodology, technology representation, and lessons learned.

cost and performance↗

Control design for robust stability in linear regulators: Application to aerospace flight control

Time domain stability robustness analysis and design for linear multivariable uncertain systems with bounded uncertainties is the central theme of the research. After reviewing the recently developed upper bounds on the linear elemental (structured), time varying perturbation of an asymptotically stable linear time invariant regulator, it is shown that it is possible to further improve these bounds by employing state transformations. Then introducing a quantitative measure called the stability robustness index, a state feedback conrol design algorithm is presented for a general linear regulator problem and then specialized to the case of modal systems as well as matched systems. The extension of the algorithm to stochastic systems with Kalman filter as the state estimator is presented. Finally an algorithm for robust dynamic compensator design is presented using Parameter Optimization (PO) procedure. Applications in a aircraft control and flexible structure control are presented along with a comparison with other existing methods.

Yedavalli, R. K.↗