Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Simulation methods.”

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 703 records · Page 39

Simulation of Plasma Emission in Magnetized Plasmas

The recent Parker Solar Probe observations of type III radio bursts show that the effects of the finite background magnetic field can be an important factor in the interpretation of data. In the present paper, the effects of the background magnetic field on the plasma-emission process, which is believed to be the main emission mechanism for solar coronal and interplanetary type III radio bursts, are investigated by means of the particle-in-cell simulation method. The effects of the ambient magnetic field are systematically surveyed by varying the ratio of plasma frequency to electron gyrofrequency. The present study shows that for a sufficiently strong ambient magnetic field, the wave–particle interaction processes lead to a highly field-aligned longitudinal mode excitation and anisotropic electron velocity distribution function, accompanied by a significantly enhanced plasma emission at the second-harmonic plasma frequency. For such a case, the polarization of the harmonic emission is almost entirely in the sense of extraordinary mode. On the other hand, for moderate strengths of the ambient magnetic field, the interpretation of the simulation result is less clear. The underlying nonlinear-mode coupling processes indicate that to properly understand and interpret the simulation results requires sophisticated analyses involving interactions among magnetized plasma normal modes, including the two transverse modes of the magneto-active plasma, namely, the extraordinary and ordinary modes, as well as electron-cyclotron-whistler, plasma oscillation, and upper-hybrid modes. At present, a nonlinear theory suitable for quantitatively analyzing such complex-mode coupling processes in magnetized plasmas is incomplete, which calls for further theoretical research, but the present simulation results could provide a guide for future theoretical efforts.

79 ASTRONOMY AND ASTROPHYSICS↗

Chapter 8: A 'Global' View on Bifacial Gain: Dependence on Geographic Location and Environmental Conditions

The simulation results presented in this chapter demonstrate how bifacial PV systems can bring a performance gain all around the globe, resolving a certain misbelief that they may only form a niche market restricted to specific locations or installations. Global predictions, based on irradiance, indicate bifacial gains in the order of 50% of the natural albedo for equator-facing systems with optimized tilt angle. In locations with low clearness index, the bifacial gain can even be higher. At system level the bifacial gain will be reduced by 20% to 50% relative to a free-standing panel, due to increased self-shading and to limitation of the diffuse irradiance on the rear by adjacent sheds. Even though the rear side irradiance of a bifacial module is inherently inhomogeneous, due to self-shading and external mounting structures, our simulations have shown that the RSD of the total irradiance does not exceed 5%, even at very high albedo, and will usually be much smaller. The simulation methods applied in this chapter and their ongoing integration into commercially available software tools, like PVsyst, will give system designers and installers more visibility on the configuration layout of a bifacial PV plant at their specific location, the associated bifacial gain and the effect on their return on investment.

14 SOLAR ENERGY↗

Designing, synthesizing, and modeling active fluids

We review recent advances in the design, synthesis, and modeling of active fluids. Active fluids have been at the center of many technological innovations and theoretical advances over the past two decades. Research on this new class of fluids has been inspired by the fascinating and remarkably efficient strategies that biological systems employ, leading to the development of biomimetic nano- and micro-machines and swimmers. The review encompasses active fluids on both the nano- and micro-scale. We start with examples of biological active systems before we discuss how experimentalists leverage novel propulsion mechanisms to power nano- and micro-machines. Furthermore, we then examine how the study of these far-from-equilibrium systems has prompted the development of new simulation methods and theoretical models in nonequilibrium physics to account for their mechanical, thermodynamic, and emergent properties. Recent advances in the field have paved the way for the design, synthesis, and modeling of autonomous systems at the nano- and micro-scale and opened the door to the development of soft matter robotics.

36 MATERIALS SCIENCE↗

Update on the Benchmark Analysis of the Jupiter Experiments: Plutonium Moderated by Lead

The Jupiter experiment was a series of measurements in the summer of 2017 to analyze the effects of voids in lead moderated plutonium assemblies. This experiment was part of a larger campaign that took place at the National Criticality Experiments Research Center (NCERC) in collaboration with the Japan Atomic Energy Agency (JAEA) that also investigated assemblies with differing uranium enrichments. The purpose for this campaign was to provide data useful for a potential lead-bismuth cooled accelerator driven system. This data is needed due to the low number of benchmarks involving lead. Given the wide use of lead for shielding in addition to new reactor designs, there is a need for further benchmarking that this experiment could help fulfill. In addition, adding more benchmarks to the International Criticality Safety Benchmark Evaluation Project (ICSBEP) is helpful for further validation of simulation methods used for criticality safety. This summary is an update on the current progress of this analysis, and all results shown are preliminary.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Sequential lateral solidification of silicon thin films on low-k dielectrics for low temperature integration

We present the excimer laser crystallization of amorphous silicon on a low dielectric constant (low-k) insulator for very large scale integration monolithic 3D integration and demonstrate that low dielectric constant materials are suitable substrates for 3D integration through laser crystallization of silicon thin films. We crystallized 100 nm amorphous silicon on top of SiO{sub 2} and SiCOH (low-k) dielectrics, at different material thicknesses (1 μm, 0.75 μm, and 0.5 μm). The amorphous silicon crystallization on low-k dielectric requires 35% less laser energy than on an SiO{sub 2} dielectric. This difference is related to the thermal conductivity of the two materials, in agreement with one dimensional simulations of the crystallization process. We analyzed the morphology of the material through defect-enhanced microscopy, Raman spectroscopy, and X-ray diffraction analysis. SEM micrographs show that polycrystalline silicon is characterized by micron-long grains with an average width of 543 nm for the SiO{sub 2} sample and 570 nm for the low-k samples. Comparison of the Raman spectra does not show any major difference in film quality for the two different dielectrics, and polycrystalline silicon peaks are closely placed around 517 cm{sup −1}. From X-ray diffraction analysis, the material crystallized on SiO{sub 2} shows a preferential (111) crystal orientation. In the SiCOH case, the 111 peak strength decreases dramatically and samples do not show preferential crystal orientation. A 1D finite element method simulation of the crystallization process on a back end of line structure shows that copper (Cu) damascene interconnects reach a temperature of 70 °C or lower with a 0.5 μm dielectric layer between the Cu and the molten Si layer, a favorable condition for monolithic 3D integration.

75 CONDENSED MATTER PHYSICS, SUPERCONDUCTIVITY AND↗

Shielding implications for secondary neutrons and photons produced within the patient during IMPT

Purpose: Intensity modulated proton therapy (IMPT) uses a combination of computer controlled spot scanning and spot-weight optimized planning to irradiate the tumor volume uniformly. In contrast to passive scattering systems, secondary neutrons and photons produced from inelastic proton interactions within the patient represent the major source of emitted radiation during IMPT delivery. Various published studies evaluated the shielding considerations for passive scattering systems but did not directly address secondary neutron production from IMPT and the ambient dose equivalent on surrounding occupational and nonoccupational work areas. Thus, the purpose of this study was to utilize Monte Carlo simulations to evaluate the energy and angular distributions of secondary neutrons and photons following inelastic proton interactions within a tissue-equivalent phantom for incident proton spot energies between 70 and 250 MeV.Methods: Monte Carlo simulation methods were used to calculate the ambient dose equivalent of secondary neutrons and photons produced from inelastic proton interactions in a tissue-equivalent phantom. The angular distribution of emitted neutrons and photons were scored as a function of incident proton energy throughout a spherical annulus at 1, 2, 3, 4, and 5 m from the phantom center. Appropriate dose equivalent conversion factors were applied to estimate the total ambient dose equivalent from secondary neutrons and photons.Results: A reference distance of 1 m from the center of the patient was used to evaluate the mean energy distribution of secondary neutrons and photons and the resulting ambient dose equivalent. For an incident proton spot energy of 250 MeV, the total ambient dose equivalent (3.6 Multiplication-Sign 10{sup -3} mSv per proton Gy) was greatest along the direction of the incident proton spot (0 Degree-Sign -10 Degree-Sign ) with a mean secondary neutron energy of 71.3 MeV. The dose equivalent decreased by a factor of 5 in the backward direction (170 Degree-Sign -180 Degree-Sign ) with a mean energy of 4.4 MeV. An 8 Multiplication-Sign 8 Multiplication-Sign 8 cm{sup 3} volumetric spot distribution (5 mm FWHM spot size, 4 mm spot spacing) optimized to produce a uniform dose distribution results in an ambient dose equivalent of 4.5 Multiplication-Sign 10{sup -2} mSv per proton Gy in the forward direction.Conclusions: This work evaluated the secondary neutron and photon emission due to monoenergetic proton spots between 70 and 250 MeV, incident on a tissue equivalent phantom. Example calculations were performed to estimate concrete shield thickness based upon appropriate workload and shielding design assumptions. Although lower than traditional passive scattered proton therapy systems, the ambient dose equivalent from secondary neutrons produced by the patient during IMPT can be significant relative to occupational and nonoccupational workers in the vicinity of the treatment vault. This work demonstrates that Monte Carlo simulations are useful as an initial planning tool for studying the impact of the treatment room and maze design on surrounding occupational and nonoccupational work areas.

61 RADIATION PROTECTION AND DOSIMETRY↗

ADVANCED SEISMIC BASE ISOLATION METHODS FOR MODULAR REACTORS

Advanced technologies for structural design and construction have the potential for major impact not only on nuclear power plant construction time and cost, but also on the design process and on the safety, security and reliability of next generation of nuclear power plants. In future Generation IV (Gen IV) reactors, structural and seismic design should be much more closely integrated with the design of nuclear and industrial safety systems, physical security systems, and international safeguards systems. Overall reliability will be increased, through the use of replaceable and modular equipment, and through design to facilitate on-line monitoring, in-service inspection, maintenance, replacement, and decommissioning. Economics will also receive high design priority, through integrated engineering efforts to optimize building arrangements to minimize building heights and footprints. Finally, the licensing approach will be transformed by becoming increasingly performance based and technology neutral, using best-estimate simulation methods with uncertainty and margin quantification. In this context, two structural engineering technologies, seismic base isolation and modular steel-plate/concrete composite structural walls, are investigated. These technologies have major potential to (1) enable standardized reactor designs to be deployed across a wider range of sites, (2) reduce the impact of uncertainties related to site-specific seismic conditions, and (3) alleviate reactor equipment qualification requirements. For Gen IV reactors the potential for deliberate crashes of large aircraft must also be considered in design. This report concludes that base-isolated structures should be decoupled from the reactor external event exclusion system. As an example, a scoping analysis is performed for a rectangular, decoupled external event shell designed as a grillage. This report also reviews modular construction technology, particularly steel-plate/concrete construction using factory prefabricated structural modules, for application to external event shell and base isolated structures.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Proton-induced x-ray fluorescence CT imaging

Purpose: To demonstrate the feasibility of proton-induced x-ray fluorescence CT (pXFCT) imaging of gold in a small animal sized object by means of experiments and Monte Carlo (MC) simulations. Methods: First, proton-induced gold x-ray fluorescence (pXRF) was measured as a function of gold concentration. Vials of 2.2 cm in diameter filled with 0%–5% Au solutions were irradiated with a 220 MeV proton beam and x-ray fluorescence induced by the interaction of protons, and Au was detected with a 3 × 3 mm{sup 2} CdTe detector placed at 90° with respect to the incident proton beam at a distance of 45 cm from the vials. Second, a 7-cm diameter water phantom containing three 2.2-diameter vials with 3%–5% Au solutions was imaged with a 7-mm FWHM 220 MeV proton beam in a first generation CT scanning geometry. X-rays scattered perpendicular to the incident proton beam were acquired with the CdTe detector placed at 45 cm from the phantom positioned on a translation/rotation stage. Twenty one translational steps spaced by 3 mm at each of 36 projection angles spaced by 10° were acquired, and pXFCT images of the phantom were reconstructed with filtered back projection. A simplified geometry of the experimental data acquisition setup was modeled with the MC TOPAS code, and simulation results were compared to the experimental data. Results: A linear relationship between gold pXRF and gold concentration was observed in both experimental and MC simulation data (R{sup 2} > 0.99). All Au vials were apparent in the experimental and simulated pXFCT images. Specifically, the 3% Au vial was detectable in the experimental [contrast-to-noise ratio (CNR) = 5.8] and simulated (CNR = 11.5) pXFCT image. Due to fluorescence x-ray attenuation in the higher concentration vials, the 4% and 5% Au contrast were underestimated by 10% and 15%, respectively, in both the experimental and simulated pXFCT images. Conclusions: Proton-induced x-ray fluorescence CT imaging of 3%–5% gold solutions in a small animal sized water phantom has been demonstrated for the first time by means of experiments and MC simulations.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Operation of a semiconductor opening switch at ultrahigh current densities

The operation of a semiconductor opening switch (SOS diode) at cutoff current densities of tens of kA/cm{sup 2} is studied. In experiments, the maximum reverse current density reached 43 kA/cm{sup 2} for {approx}40 ns. Experimental data on SOS diodes with a p{sup +}-p-n-n{sup +} structure and a p-n junction depth from 145 to 180 {mu}m are presented. The dynamics of electron-hole plasma in the diode at pumping and current cutoff stages is studied by numerical simulation methods. It is shown that current cutoff is associated with the formation of an electric field region in a thin ({approx}45 {mu}m) layer of the structure's heavily doped p-region, in which the acceptor concentration exceeds 10{sup 16} cm{sup -3}, and the current cutoff process depends weakly on the p-n junction depth.

36 MATERIALS SCIENCE↗

Direct comparison between experiments and dislocation dynamics simulations of high rate deformation of single crystal copper

A long standing challenge in computational materials science is to establish a quantitative connection between the macroscopic properties of plastic deformation with the microscopic mechanisms of dislocations in crystalline materials. Although the discrete dislocation dynamics (DDD) simulation method has been developed for several decades with the goal of addressing this challenge, a one-to-one comparison between the DDD predictions on single crystal stress–strain curves and experimental measurements under identical conditions has not been possible to date. Such a comparison is an essential step towards establishing a dislocation-physics based theory of plasticity and a multiscale framework of the plastic behaviors of crystalline materials. Here we provide direct comparisons between the stress–strain curves of Cu single crystals under high strain rate loading in the [0 0 1] and [0 1 1] directions obtained from miniaturized desktop Kolsky bar experiments and those from DDD simulations under identical loading conditions. With an appropriate set of parameters, DDD simulations can produce stress–strain curves that are in reasonable agreement with the experimental results. However, the dislocation mobility values needed to achieve this agreement are an order of magnitude lower than expected based on previous measurements and atomistic simulations. We hypothesize that this discrepancy could be caused by drag forces from jogs and point defects produced during the plastic deformation. Cross-slip of screw dislocations is also found to be necessary to capture the experimental stress–strain behavior, especially for the [0 1 1] loading direction. Finally, this work provides an example of how direct comparisons between DDD simulations and experimental measurements can provide new insight into the fundamental mechanisms of plastic deformation.

36 MATERIALS SCIENCE↗

Modeling the 4D discharge of lithium-ion batteries with a multiscale time-dependent deep learning framework

The lithium-ion battery (LIB) field is moving towards the direction of investigating spatially resolved physical phenomena in the 3D porous microstructure of electrodes. These pore-scale simulations give new insights into the local dynamics of lithiation/de-lithiation and charge transport, Nevertheless, the computational time of these simulations limits the integration of these models in optimization workflows of cycling conditions or electrode manufacturing processes. Machine learning models present a way of assessing in real-time the performance of materials. While several successful techniques for replicating simulations with machine learning have been proposed, this case study presents a more demanding problem, due to the necessity of understanding the behavior of heterogeneous 3D local data, as it evolves in time: this poses both a scientific and a technical challenge. To this end, we propose an autoregressive multiscale convolutional neural network model to predict relevant quantities at the pore-scale in the solid phase: the lithium concentration (in the active material) and potential (in the active material and carbon binder). Here, these are ultimately used to reconstruct the battery discharge curve. 3D images of the electrode microstructures are the input to the network, trained with a dataset of finite element method simulations to predict the discharge behavior of the cathode side in lithium ion batteries. We propose this machine learning model as a proof-of-concept of the applicability of multiscale networks for time-dependent physics problems. The trained model exhibits very high accuracy (with errors lower than 2 %) in forecasting the discharge behavior of new unseen cathodes.

25 ENERGY STORAGE↗

Development of a machine learning model for polyethylene pyrolysis using a detailed reaction mechanism

Waste plastics have recently received significant attention as the issue of waste generation continues to increase. Thermal conversion processes, such as pyrolysis and gasification, are attractive potential technologies for utilizing waste plastics and reducing overall waste generation. Efficient utilization of plastics requires a detailed understanding of the conversion process such as pyrolysis and gasification. However, a mechanistic understanding of these processes lead to large and complex kinetic schemes that are not suited for large-scale and long-time simulation methods. Currently, most modeling approaches for pyrolysis and gasification rely on globally lumped, simplified kinetic schemes that provide results that are classified by their product type and not individual species, which limit the level of fidelity achieved via modeling. A machine learning (ML) model has been developed for the primary reactions of high-density polyethylene (HDPE) in an attempt to increase computational efficiency while still maintaining a high level of detail and accuracy. The ML model is trained on a detailed reaction mechanism containing 42 total species and 737 chemical reactions. A DeepONet branch and trunk architecture was adopted to train the model using time-steps relevant to computational fluid dynamics simulations. The ML used physics-informed loss functions to ensure mass conservation. The surrogate model has been deployed in simple MFiX CFD simulations, single particle and an experimental drop tube reactor, and has shown promising performance compared to the original scheme.

Houston, Ross↗

Improving deep learning performance for predicting large-scale geological ${{CO}_{2}}$ sequestration modeling through feature coarsening

Physics-based reservoir simulation for fluid flow in porous media is a numerical simulation method to predict the temporal-spatial patterns of state variables (e.g. pressure p) in porous media, and usually requires prohibitively high computational expense due to its non-linearity and the large number of degrees of freedom (DoF). This work describes a deep learning (DL) workflow to predict the pressure evolution as fluid flows in large-scale 3-dimensional(3D) heterogeneous porous media. In particular, we develop an efficient feature coarsening technique to extract the most representative information and perform the training and prediction of DL at the coarse scale, and further recover the resolution at the fine scale by spatial interpolation. We validate the DL approach to predict pressure field against physics-based simulation data for a field-scale 3D geologic CO 2 sequestration reservoir model. We evaluate the impact of feature coarsening on DL performance, and observe that the feature coarsening not only decreases the training time by >74% and reduces the memory consumption by >75%, but also maintains temporal error 0.63% on average. Besides, the DL workflow provides predictive efficiency with 1406 times speedup compared to physics-based numerical simulation. The key findings from this research significantly improve the training and prediction efficiency of deep learning model to deal with large-scale heterogeneous reservoir models, and thus it can also be further applied to accelerate workflows of history matching and reservoir optimization for close-loop reservoir management.

58 GEOSCIENCES↗

Overall results of and lessons learned from the IAEA CRP on sodium natural circulation test performed during the Phenix end-of-life experiments

In 2007, the International Atomic Energy Agency (IAEA) launched the Coordinated Research Project (CRP) 'Control Rod Withdrawal and Sodium Natural Circulation Tests Performed during the Phenix End-of-Life Experiments'. The overall purpose of the CRP, performed within the framework of the IAEA programme in support of innovative fast reactor technology development and deployment, is to improve the Member States' analytical capabilities in the various fields of research and design of sodium-cooled fast reactors through data and codes verification and validation. In particular the CRP, taking advantage of the End-of-Life set of experiments performed before the final shut-down of the French prototype fast breeder power reactor Phenix, aims at improving fast reactor simulation methods and design capabilities in the field of temperature and power distribution evaluation, as well as of the analysis of sodium natural circulation phenomena. The paper presents the overall results of the CRP, including blind calculations and post-test and sensitivity analyses carried out by the CRP participants, as well as lessons learned and recommendations for further future implementations to resolve open issues. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Simulation and quasilinear theory of proton firehose instability

The electromagnetic proton firehose instability is driven by excessive parallel temperature anisotropy, T{sub ∥} > T{sub ⊥} (or more precisely, parallel pressure anisotropy, P{sub ∥} > P{sub ⊥}) in high-beta plasmas. Together with kinetic instabilities driven by excessive perpendicular temperature anisotropy, namely, electromagnetic proton cyclotron and mirror instabilities, its role in providing the upper limit for the temperature anisotropy in the solar wind is well-known. A recent Letter [Seough et al., Phys. Rev. Lett. 110, 071103 (2013)] employed quasilinear kinetic theory for these instabilities to explain the observed temperature anisotropy upper bound in the solar wind. However, the validity of quasilinear approach has not been rigorously tested until recently. In a recent paper [Seough et al., Phys. Plasmas 21, 062118 (2014)], a comparative study is carried out for the first time in which quasilinear theory of proton cyclotron instability is tested against results obtained from the particle-in-cell simulation method, and it was demonstrated that the agreement was rather excellent. The present paper addresses the same issue involving the proton firehose instability. Unlike the proton cyclotron instability, however, it is found that the quasilinear approximation enjoys only a limited range of validity, especially for the wave dynamics and for the relatively high-beta regime. Possible causes and mechanisms responsible for the discrepancies are speculated and discussed.

43 PARTICLE ACCELERATORS↗

A practical approach to the sensitivity analysis for kinetic Monte Carlo simulation of heterogeneous catalysis

Lattice kinetic Monte Carlo simulations have become a vital tool for predictive quality atomistic understanding of complex surface chemical reaction kinetics over a wide range of reaction conditions. In order to expand their practical value in terms of giving guidelines for atomic level design of catalytic systems, it is very desirable to readily evaluate a sensitivity analysis for a given model. The result of such a sensitivity analysis quantitatively expresses the dependency of the turnover frequency, being the main output variable, on the rate constants entering the model. In the past the application of sensitivity analysis, such as Degree of Rate Control, has been hampered by its exuberant computational effort required to accurately sample numerical derivatives of a property that is obtained from a stochastic simulation method. Here in this study we present an efficient and robust three stage approach that is capable of reliably evaluating the sensitivity measures for stiff microkinetic models as we demonstrate using CO oxidation on RuO 2 (110) as a prototypical reaction. In a first step, we utilize the Fisher Information Matrix for filtering out elementary processes which only yield negligible sensitivity. Then we employ an estimator based on linear response theory for calculating the sensitivity measure for non-critical conditions which covers the majority of cases. Finally we adopt a method for sampling coupled finite differences for evaluating the sensitivity measure of lattice based models. This allows efficient evaluation even in critical regions near a second order phase transition that are hitherto difficult to control. The combined approach leads to significant computational savings over straightforward numerical derivatives and should aid in accelerating the nano scale design of heterogeneous catalysts.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Delocalized Excitation Transfer in Open Quantum Systems with Long-Range Interactions

The interplay between coherence and system-environment interactions is at the basis of a wide range of phenomena, from quantum information processing to charge and energy transfer in molecular systems, biomolecules, and photochemical materials. In this work, we use a Frenkel exciton model with long-range interacting qubits coupled to a damped collective bosonic mode to investigate vibrationally assisted transfer processes in donor-acceptor systems featuring internal substructures analogous to light-harvesting complexes. We find that certain delocalized excitonic states maximize the transfer rate and that the entanglement is preserved during the dissipative transfer over a wide range of parameters. We investigate the reduction in transfer caused by static disorder, white noise, and finite temperature and study how transfer efficiency scales as a function of the number of dimerized monomers and the component number of each monomer, finding which excitonic states lead to optimal transfer. Finally, we provide a realistic experimental setting to realize this model in analog trapped-ion quantum simulators. Analog quantum simulation of systems comprising many and increasingly complex monomers could offer valuable insights into the design of light-harvesting materials, particularly in the nonperturbative intermediate parameter regime examined in this study, where classical simulation methods are resource intensive.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

Annealed Importance Sampling Reversible Jump MCMC algorithms

It will soon be 20 years since reversible jump Markov chain Monte Carlo (RJ-MCMC) algorithms have been proposed. They have significantly extended the scope of Markov chain Monte Carlo simulation methods, offering the promise to be able to routinely tackle transdimensional sampling problems, as encountered in Bayesian model selection problems for example, in a principled and flexible fashion. Their practical efficient implementation, however, still remains a challenge. A particular difficulty encountered in practice is in the choice of the dimension matching variables (both their nature and their distribution) and the reversible transformations which allow one to define the one-to-one mappings underpinning the design of these algorithms. Indeed, even seemingly sensible choices can lead to algorithms with very poor performance. The focus of this paper is the development and performance evaluation of a method, annealed importance sampling RJ-MCMC (aisRJ), which addresses this problem by mitigating the sensitivity of RJ-MCMC algorithms to the aforementioned poor design. As we shall see the algorithm can be understood as being an “exact approximation” of an idealized MCMC algorithm that would sample from the model probabilities directly in a model selection set-up. Such an idealized algorithm may have good theoretical convergence properties, but typically cannot be implemented, and our algorithms can approximate the performance of such idealized algorithms to an arbitrary degree while not introducing any bias for any degree of approximation. Our approach combines the dimension matching ideas of RJ-MCMC with annealed importance sampling and its Markov chain Monte Carlo implementation. We illustrate the performance of the algorithm with numerical simulations which indicate that, although the approach may at first appear computationally involved, it is in fact competitive.

annealed importance sampling↗