skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Girsanov's transformation based variance reduced Monte Carlo simulation schemes for reliability estimation in nonlinear stochastic dynamics

Abstract

The study considers the problem of simulation based time variant reliability analysis of nonlinear randomly excited dynamical systems. Attention is focused on importance sampling strategies based on the application of Girsanov's transformation method. Controls which minimize the distance function, as in the first order reliability method (FORM), are shown to minimize a bound on the sampling variance of the estimator for the probability of failure. Two schemes based on the application of calculus of variations for selecting control signals are proposed: the first obtains the control force as the solution of a two-point nonlinear boundary value problem, and, the second explores the application of the Volterra series in characterizing the controls. The relative merits of these schemes, vis-à-vis the method based on ideas from the FORM, are discussed. Illustrative examples, involving archetypal single degree of freedom (dof) nonlinear oscillators, and a multi-degree of freedom nonlinear dynamical system, are presented. The credentials of the proposed procedures are established by comparing the solutions with pertinent results from direct Monte Carlo simulations. - Highlights: • The distance minimizing control forces minimize a bound on the sampling variance. • Establishing Girsanov controls via solution of a two-point boundary value problem. • Girsanov controls viamore » Volterra's series representation for the transfer functions.« less

Authors:
;
Publication Date:
OSTI Identifier:
22622308
Resource Type:
Journal Article
Resource Relation:
Journal Name: Journal of Computational Physics; Journal Volume: 341; Other Information: Copyright (c) 2017 Elsevier Science B.V., Amsterdam, The Netherlands, All rights reserved.; Country of input: International Atomic Energy Agency (IAEA)
Country of Publication:
United States
Language:
English
Subject:
97 MATHEMATICAL METHODS AND COMPUTING; BOUNDARY-VALUE PROBLEMS; COMPARATIVE EVALUATIONS; COMPUTERIZED SIMULATION; DEGREES OF FREEDOM; FAILURES; MATHEMATICAL SOLUTIONS; MONTE CARLO METHOD; NONLINEAR PROBLEMS; OSCILLATORS; PROBABILITY; RANDOMNESS; REDUCTION; RELIABILITY; SAMPLING; STOCHASTIC PROCESSES; TRANSFER FUNCTIONS; TRANSFORMATIONS

Citation Formats

Kanjilal, Oindrila, E-mail: oindrila@civil.iisc.ernet.in, and Manohar, C.S., E-mail: manohar@civil.iisc.ernet.in. Girsanov's transformation based variance reduced Monte Carlo simulation schemes for reliability estimation in nonlinear stochastic dynamics. United States: N. p., 2017. Web. doi:10.1016/J.JCP.2017.03.047.
Kanjilal, Oindrila, E-mail: oindrila@civil.iisc.ernet.in, & Manohar, C.S., E-mail: manohar@civil.iisc.ernet.in. Girsanov's transformation based variance reduced Monte Carlo simulation schemes for reliability estimation in nonlinear stochastic dynamics. United States. doi:10.1016/J.JCP.2017.03.047.
Kanjilal, Oindrila, E-mail: oindrila@civil.iisc.ernet.in, and Manohar, C.S., E-mail: manohar@civil.iisc.ernet.in. 2017. "Girsanov's transformation based variance reduced Monte Carlo simulation schemes for reliability estimation in nonlinear stochastic dynamics". United States. doi:10.1016/J.JCP.2017.03.047.
@article{osti_22622308,
title = {Girsanov's transformation based variance reduced Monte Carlo simulation schemes for reliability estimation in nonlinear stochastic dynamics},
author = {Kanjilal, Oindrila, E-mail: oindrila@civil.iisc.ernet.in and Manohar, C.S., E-mail: manohar@civil.iisc.ernet.in},
abstractNote = {The study considers the problem of simulation based time variant reliability analysis of nonlinear randomly excited dynamical systems. Attention is focused on importance sampling strategies based on the application of Girsanov's transformation method. Controls which minimize the distance function, as in the first order reliability method (FORM), are shown to minimize a bound on the sampling variance of the estimator for the probability of failure. Two schemes based on the application of calculus of variations for selecting control signals are proposed: the first obtains the control force as the solution of a two-point nonlinear boundary value problem, and, the second explores the application of the Volterra series in characterizing the controls. The relative merits of these schemes, vis-à-vis the method based on ideas from the FORM, are discussed. Illustrative examples, involving archetypal single degree of freedom (dof) nonlinear oscillators, and a multi-degree of freedom nonlinear dynamical system, are presented. The credentials of the proposed procedures are established by comparing the solutions with pertinent results from direct Monte Carlo simulations. - Highlights: • The distance minimizing control forces minimize a bound on the sampling variance. • Establishing Girsanov controls via solution of a two-point boundary value problem. • Girsanov controls via Volterra's series representation for the transfer functions.},
doi = {10.1016/J.JCP.2017.03.047},
journal = {Journal of Computational Physics},
number = ,
volume = 341,
place = {United States},
year = 2017,
month = 7
}
  • Recently, Fokker–Planck based particle Monte Carlo schemes have been proposed and evaluated for simulations of rarefied gas flows [1–3]. In this paper, the variance reduction for particle Monte Carlo simulations based on the Fokker–Planck model is considered. First, deviational based schemes were derived and reviewed, and it is shown that these deviational methods are not appropriate for practical Fokker–Planck based rarefied gas flow simulations. This is due to the fact that the deviational schemes considered in this study lead either to instabilities in the case of two-weight methods or to large statistical errors if the direct sampling method is applied.more » Motivated by this conclusion, we developed a novel scheme based on correlated stochastic processes. The main idea here is to synthesize an additional stochastic process with a known solution, which is simultaneously solved together with the main one. By correlating the two processes, the statistical errors can dramatically be reduced; especially for low Mach numbers. To assess the methods, homogeneous relaxation, planar Couette and lid-driven cavity flows were considered. For these test cases, it could be demonstrated that variance reduction based on parallel processes is very robust and effective.« less
  • Planning and scheduling of power system operation for future loads involves extensive studies of reliability analysis and production costing to ensure secure and economical supply of electric energy. Sensitivity analysis plays an important role in fine tuning the operating plans in the presence of various uncertainties. In this paper, the application of a recently developed technique known as Perturbation Analysis (PA) to the generation system reliability analysis using Monte Carlo simulation is presented. Specifically, the PA is used to find the derivatives of some of the reliability indices. Starting with a general introduction of the Monte Carlo simulation, the powermore » system model for the application of PA is described. This is followed by a thorough discussion of the generation and propagation of perturbations to estimate the derivatives during the chronological simulation. A numerical examples is presented in which the PA derivative estimates of a 6 unit thermal system are compared to the estimates obtained from analytical methods. The paper concludes with the suggestions for future work needed to expand the application of PA in this area.« less
  • This paper proposes a comprehensive framework for accelerating population balance-Monte Carlo (PBMC) simulation of particle coagulation dynamics. By combining Markov jump model, weighted majorant kernel and GPU (graphics processing unit) parallel computing, a significant gain in computational efficiency is achieved. The Markov jump model constructs a coagulation-rule matrix of differentially-weighted simulation particles, so as to capture the time evolution of particle size distribution with low statistical noise over the full size range and as far as possible to reduce the number of time loopings. Here three coagulation rules are highlighted and it is found that constructing appropriate coagulation rule providesmore » a route to attain the compromise between accuracy and cost of PBMC methods. Further, in order to avoid double looping over all simulation particles when considering the two-particle events (typically, particle coagulation), the weighted majorant kernel is introduced to estimate the maximum coagulation rates being used for acceptance–rejection processes by single-looping over all particles, and meanwhile the mean time-step of coagulation event is estimated by summing the coagulation kernels of rejected and accepted particle pairs. The computational load of these fast differentially-weighted PBMC simulations (based on the Markov jump model) is reduced greatly to be proportional to the number of simulation particles in a zero-dimensional system (single cell). Finally, for a spatially inhomogeneous multi-dimensional (multi-cell) simulation, the proposed fast PBMC is performed in each cell, and multiple cells are parallel processed by multi-cores on a GPU that can implement the massively threaded data-parallel tasks to obtain remarkable speedup ratio (comparing with CPU computation, the speedup ratio of GPU parallel computing is as high as 200 in a case of 100 cells with 10 000 simulation particles per cell). These accelerating approaches of PBMC are demonstrated in a physically realistic Brownian coagulation case. The computational accuracy is validated with benchmark solution of discrete-sectional method. The simulation results show that the comprehensive approach can attain very favorable improvement in cost without sacrificing computational accuracy.« less
  • The radiation transport kernel has two factors, namely, the transition kernel and the collision kernel. Using the modified moments equations for analog and nonanalog games, families of zero-variance biasing schemes have been obtained for the two factors for evaluating reaction rate integrals using last-event, collision, and partial score estimators. It is shown that all zero-variance biasing schemes available in the literature belong to the respective family for each estimator.
  • A zero variance path-stretching biasing scheme proposed for a special case by Dwivedi is derived in full generality. The procedure turns out to be the generalization of the exponential transform. It is shown that the biased game can be interpreted as an analog simulation procedure, thus saving some computational effort in comparison with the corresponding nonanalog game.