Search NASA⌕ Search

SEARCH · Search NASA

Results for “Fast reactor”

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 199 records · Page 11

Enhanced Performance Fast Reactors with Engineered Passive Safety Systems

This report summarizes the work and results of NEUP Project #15-8251, "Enhanced Performance Fast Reactors with Engineered Passive Safety System," performed from 2015-2018 at UC Berkeley and Argonne National Laboratory. This project concerns the incorporation of a novel engineered passive safety system, called the Autonomous Reactivity Control (ARC) system, into Sodium-cooled Fast Reactors (SFRs) to improve their transient and economic performance. The first part of this project consisted of fleshing out the basic principles and the manufacturing process for an SFR fuel assembly with the ARC system installed, with the details provided in high-quality, step-by-step CAD drawings. Once the detailed design is laid out, high-fidelity simulations were performed to optimize the dimensions of important ARC system components and to quantify the standalone time-response of ARC systems in a typical SFR environment. These high-fidelity simulations were used to create reduced-order models of ARC system response which could be incorporated into the SAS4A/SASSYS-1 whole-core transient models. In order to accurately account for the time-response of ARC systems in whole-core transient simulations, modifications were made to the SAS4A/SASSYS-1 transient code. These new capabilities were tested and incorporated into the production version of the code, and have been used extensively in this analysis. Before performing whole-core transient calculations, the two reference cores selected for this study to cover a wide design space were simulated and characterized. The two reference cores, (1) a medium-sized oxide-fueled Advanced Burner Reactor and (2) a large metal-fueled Breed-and-Burn core, challenge conventional means of achieving passive safety. Because the two cores vary substantially, simulating the ARC system in both allows for conclusions to be drawn about the applicability of ARC systems in different SFRs. Once both cores were fully characterized, transient calculations were performed to understand the complicated interplay between the ARC system and other inherent reactivity feedback mechanisms. By performing parametric studies of the transient behavior, ARC system designs that best improve transient performance in each core were identified. Once transient improvements were established, the transient calculations were extended to improve reactor economics through design modifications. Upon establishing the potential benefits of ARC system inclusion, the study examines the negative neutronic implications of incorporating ARC systems into the reference cores. The study concludes by discussing a number of knowledge gaps that exist which may have important implications for ARC system implementation. Ultimately, the results of this project indicate that ARC systems may provide an effective means to improve transient performance in certain types of systems. The transient simulations in the oxide Advanced Burner Reactor showed that significant margin to coolant boiling and fuel melting could be gained through the use of a properly designed ARC system, thus improving inherent safety. While oscillatory behavior was induced in the transient response for certain combinations of ARC system parameters, the oscillation-free design space that can be effectively used to improve performance was found to be large, and the additional margin was demonstrated to allow economical improvements to the core design. Conversely, difficulties were encountered in the Breed-and-Burn core due to fundamental aspects of the core design. The combination of a particularly strong positive coolant void feedback and the use of metallic fuel allowed for diverging oscillations to manifest in response to certain transient initiators, and in the end it was found that the current ARC system design could not simultaneously enable passive safety in all transients examined. In order for complete passive safety to be achieved, a mechanism for slowing the reverse actuation of ARC systems is necessary. Regardless, the standard ARC system was able to improve the performance of two out of the three examined transients by increasing the margin to boiling in Loss of Heat Sink transients and avoiding boiling in Transient Overpower accidents. Additionally, no significant barriers to implementation were found, although the issue of tritium production should be closely monitored in any core incorporating ARC systems. Still, a number of technology gaps exist, and these are discussed at the conclusion of the report to recommend directions for future work.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

An Assessment of Fission Product Scrubbing in Sodium Pools Following a Core Damage Event in a Sodium Cooled Fast Reactor

The U.S. Nuclear Regulatory Commission has stated that mechanistic source term (MST) calculations are expected to be required as part of the advanced reactor licensing process. A recent study by Argonne National Laboratory has concluded that fission product scrubbing in sodium pools is an important aspect of an MST calculation for a sodium-cooled fast reactor (SFR). To model the phenomena associated with sodium pool scrubbing, a computational tool, developed as part of the Integral Fast Reactor (IFR) program, was utilized in an MST trial calculation. This tool was developed by applying classical theories of aerosol scrubbing to the decontamination of gases produced as a result of postulated fuel pin failures during an SFR accident scenario. The model currently considers aerosol capture by Brownian diffusion, inertial deposition, and gravitational sedimentation. The effects of sodium vapour condensation on aerosol scrubbing are also treated. This paper provides details of the individual scrubbing mechanisms utilized in the IFR code as well as results from a trial mechanistic source term assessment led by Argonne National Laboratory in 2016.

Fission Product Scrubbing↗

The feasibility study of small long-life gas cooled fast reactor with mixed natural Uranium/Thorium as fuel cycle input

A conceptual design study of Gas Cooled Fast Reactors with Modified CANDLE burn-up scheme has been performed. In this study, design GCFR with Helium coolant which can be continuously operated by supplying mixed Natural Uranium/Thorium without fuel enrichment plant or fuel reprocessing plant. The active reactor cores are divided into two region, Thorium fuel region and Uranium fuel region. Each fuel core regions are subdivided into ten parts (region-1 until region-10) with the same volume in the axial direction. The fresh Natural Uranium and Thorium is initially put in region-1, after one cycle of 10 years of burn-up it is shifted to region-2 and the each region-1 is filled by fresh natural Uranium/Thorium fuel. This concept is basically applied to all regions in both cores area, i.e. shifted the core of i{sup th} region into i+1 region after the end of 10 years burn-up cycle. For the next cycles, we will add only Natural Uranium and Thorium on each region-1. The calculation results show the reactivity reached by mixed Natural Uranium/Thorium with volume ratio is 4.7:1. This reactor can results power thermal 550 MWth. After reactor start-up the operation, furthermore reactor only needs Natural Uranium/Thorium supply for continue operation along 100 years.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Optimization of multi-group cross sections for fast reactor analysis

The selection of the number of broad energy groups, collapsed broad energy group boundaries, and their associated evaluation into collapsed macroscopic cross sections from a general 238-group ENDF/B-VII library dramatically impacted the k eigenvalue for fast reactor analysis. An analysis was undertaken to assess the minimum number of energy groups that would preserve problem physics; this involved studies using the 3D deterministic transport parallel code PENTRAN, the 2D deterministic transport code SCALE6.1, the Monte Carlo based MCNP5 code, and the YGROUP cross section collapsing tool on a spatially discretized MOX fuel pin comprised of 21% PUO{sub 2}-UO{sub 2} with sodium coolant. The various cases resulted in a few hundred pcm difference between cross section libraries that included the 238 multi-group reference, and cross sections rendered using various reaction and adjoint weighted cross sections rendered by the YGROUP tool, and a reference continuous energy MCNP case. Particular emphasis was placed on the higher energies characteristic of fission neutrons in a fast spectrum; adjoint computations were performed to determine the average per-group adjoint fission importance for the MOX fuel pin. This study concluded that at least 10 energy groups for neutron transport calculations are required to accurately predict the eigenvalue for a fast reactor system to within 250 pcm of the 238 group case. In addition, the cross section collapsing/weighting schemes within YGROUP that provided a collapsed library rendering eigenvalues closest to the reference were the contribution collapsed, reaction rate weighted scheme. A brief analysis on homogenization of the MOX fuel pin is also provided, although more work is in progress in this area. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Closing nuclear fuel cycle with fast reactors: problems and prospects

The closed nuclear fuel cycle (CNFC) with fast reactors (FR) is the most promising way of nuclear energetics development because it prevents spent nuclear fuel (SNF) accumulation and minimizes radwaste volume due to minor actinides (MA) transmutation. CNFC with FR requires the elaboration of safety, environmentally acceptable and economically effective methods of treatment of SNF with high burn-up and low cooling time. The up-to-date industrially implemented SNF reprocessing technologies based on hydrometallurgical methods are not suitable for the reprocessing of SNF with high burn-up and low cooling time. The alternative dry methods (such as electrorefining in molten salts or fluoride technologies) applicable for such SNF reprocessing have not found implementation at industrial scale. So the cost of SNF reprocessing by means of dry technologies can hardly be estimated. Another problem of dry technologies is the recovery of fissionable materials pure enough for dense fuel fabrication. A combination of technical solutions performed with hydrometallurgical and dry technologies (pyro-technology) is proposed and it appears to be a promising way for the elaboration of economically, ecologically and socially accepted technology of FR SNF management. This paper deals with discussion of main principle of dry and aqueous operations combination that probably would provide safety and economic efficiency of the FR SNF reprocessing. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Assembly Bowing Reactivity Calculation Methodology Applied to Lead Fast Reactor

Ducted assemblies bow during operation due to power and temperature gradients which can be influenced by operating flow rates. For fast spectrum cores using ducted assemblies, the bowing behavior follows that of the duct and because there are gaps between the ducts, the bowing can result in compaction or expansion of the active core. This local displacement can have a positive or negative impact and knowing the net effect during transients is important for system reactivity control. Keeping the net bowing reactivity worth low is possible with attentive placement of load pads above the active core and selecting load pad gap thicknesses that result in a desired behavior at standard operating conditions. This paper considers a Lead Fast Reactor (LFR) concept fueled by HALEU UO2 developed by Westinghouse Electric Company (WEC) and applies a workflow of Argonne-developed codes to estimate the core bowing reactivity worth. Using orifice flow rates grouped by assembly type, the net reactivity impact due to core assembly bowing for the LFR was found to be small and in line with other liquid metal fast reactors: +29/+32/+35 pcm, or about +$0.049/+$0.055/+$0.059, for BOEC/MOEC/EOEC, respectively.

bowing reactivity↗

The Industrial Sodium Cooled Fast Reactor

This paper investigates the use of enrichment and moderator zoning methods for optimizing the r-z power distribution within sodium cooled fast reactors. These methods allow overall greater fuel utilization in the core resulting in more fuel being irradiated near the maximum allowed thermal power. The peak-to-average power density was held to 1.18. This core design, in conjunction with a multiple-reheat Brayton power conversion system, has merit for producing an industrial level of electrical output (2400MWth, 1000MWe) from a relatively compact core size. The total core radius, including reflectors and shields, was held to 1.78m. Preliminary safety analysis suggests that positive reactivity insertion resulting from a leak between the sodium primary loop and helium power conversion system can be mitigated using simple gas-liquid centripetal separation strategies in the plant’s primary loop.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Experimental Studies of Coolant Temperature and Velocity on an Integrated Water Model of a Fast Reactor in Different Operating Regimes

The results of experimental studies of a next-generation fast reactor performed using an integrated water model (scale 1:10) are presented: local velocity over height and radius of the top chamber in a plane in a direction from core center to the intermediate heat exchangers and coolant temperature in the top chamber and other elements of the circulation loop with forced circulation and emergency cool-down by natural convection. The data were obtained using a specially designed system for performing measurements which is incorporated in the bench and provides high accuracy and rate of acquisition. Stable temperature stratification of the coolant is observed in the top, cold, and pressure chambers of the reactor, elevator enclosure, and reactor vessel cooling system as well as at the exit from the intermediate and autonomous heat exchangers in different operating regimes.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Effectiveness of actinide management in fast reactors over a finite recycling period - 15123

Long-lived nuclear waste primarily consists of transuranic (TRU) isotopes. Full recycling of TRU isotopes can, in theory, lead to a reduction in repository radiotoxicity to reference levels in as little as ∼500 years provided reprocessing and fuel fabrication losses are limited. However, over a limited time frame, the radiotoxicity of the 'final' core can dominate over reprocessing losses, leading to a much lower reduction in radiotoxicity compared to that achievable at equilibrium. In this paper, the fuel cycle code ORION is used to model the recycle of light water reactor (LWR)-produced TRUs in sodium-cooled fast reactors (SFRs). With full actinide recycling, at least 6 generations of SFRs are required in a gradual phase-out of nuclear power to achieve transmutation performance approaching the theoretical equilibrium performance. Break-even SFRs are much less effective from a point of view of reducing waste radiotoxicity. Thorium (Th)-burner fuel cycles are found to result in slightly lower radiotoxicity than U-burner fuel cycles for the same number of generations. Closed break-even Th fuel cycles require ∼3 generations of operation before their waste radiotoxicity benefits result in decay to the reference level in ∼1000 years, provided that Pa is recycled. While this is a very long time frame, it is substantially better than the time frame required by other strategies. The integrated decay heat over the scenario time frame is substantially higher for break-even Th-SFRs than break-even U-SFRs as a result of much higher {sup 90}Sr production. As decay heat affects repository sizing, this weakens the argument for the Th cycle. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

One pass core design of a super fast reactor

One pass core design for Supercritical-pressure light water-cooled fast reactor (Super FR) is proposed. The whole core is cooled with upward flow in one through flow pattern like PWR. Compared with the previous two pass core design; this new flow pattern can significantly simplify the core concept. Upper core structure, coolant flow scheme as well as refueling procedure are as simple as in PWR. In one pass core design, supercritical-pressure water is at approximately 25.0 MPa and enters the core at 280 C. degrees and is heated up in one through flow pattern upwardly to the average outlet temperature of 500 C. degrees. Great density change in vertical direction can cause significant axial power offset during the cycle. Meanwhile, Pu accumulated in the UO{sub 2} fuel blanket assemblies also introduces great power increase during cycle, which requires large amount of flow for heat removal and makes the outlet temperature of blanket low at the beginning of equilibrium cycle (BOEC). To deal with these issues, some MOX fuel is applied in the bottom region of the blanket assembly. This can help to mitigate the power change in blanket due to Pu accumulation and to increase the outlet temperature of the blanket during cycle. Neutron transport and thermohydraulics coupled calculation shows that this design can satisfy the requirement in the Super FR principle for both 500 C. degrees outlet temperature and negative coolant void reactivity. (authors)

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

Studying the effect of the upper internal structure design on thermal stratification in sodium fast reactors

Thermal stratification and mixing in large reactor pools can pose significant issues with regard to operation and safety of nuclear plants. While significant work has been done in this area on light water reactors it is more critical for sodium reactors and there is far less data with regard to low Prandtl number fluids. Here this paper focuses on the effect of the upper internal structure on the formation of thermal stratification in a scaled sodium fast reactor upper plenum. Six different upper internal structure configurations were tested to observe the effect on thermal stratification. The degree of stratification was quantified using the stratification number. Each configuration was tested from low to high flow rates at a temperature difference of 50 and 100 °C. Over this broad array of testing, it was observed that higher flow rates, more porosity in the upper internal structure design, and a smaller temperature difference yielded a lower degree of stratification. This information can be used to aid in the design of the upper internal structure in a sodium fast reactor. Furthermore, this high resolved temperature data can be used to verify models under prototypic low Prandtl numbers.

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Laser-spectroscopy testbed for impurity monitoring in liquid metal-cooled fast reactors

A significant challenge in sodium-cooled fast reactors is controlling impurities, in particular oxygen impurities, within the sodium coolant, as they can accelerate corrosion and indicate leaks. Optical methods offer the potential to rapidly detect small concentrations of both gaseous and metal impurities that accelerate corrosion, plug coolant channels, and lead to increased activation of isotopes in the coolant. We present the design and performance of an apparatus designed to enable the application of multiple optical analytical techniques, such as laser-induced breakdown spectroscopy, to detect elemental impurities in the sodium melt with high sensitivity. Here, we experimentally demonstrate the detection of characteristic sodium and oxygen spectral lines in liquid sodium, which sets the stage for the optimization of its analytical sensitivity. A robust sensor of this type integrated with the sodium cooling loop has the potential to significantly improve the safety and operational efficiency of generation IV nuclear reactors.

47 OTHER INSTRUMENTATION↗

Optimization study of Ultra-long Cycle Fast Reactor core concept

An optimization of an Ultra-long Cycle Fast Reactor (UCFR) design with a power rating of 1000 MW(electric), UCFR-1000, has been performed. Firstly, geometric optimization has been performed in the aspects of core size and core shape in terms of thermal–hydraulic (TH) feedback. Secondly, fuel composition optimization has been performed by adopting pressurized water reactor (PWR) spent fuel (SF) as a blanket material as well as natural uranium (NU). Thirdly, thorium has been loaded in the inner core for the optimization of radial power distribution. Lastly, a small-size UCFR with a power rate of 100 MWe has been developed with optimization of maximum neutron flux and fast neutron fluence limit for a short term deployable nuclear reactor. The equivalent diameter and the height of the optimized UCFR-1000 core are 5.9 and 2.4 m, respectively, while the equivalent diameter and the height of the optimized UCFR-100 core are 4.3 and 1.0 m, respectively. The size of the optimized UCFR-1000 has been enlarged in the radial direction and shortened in the axial direction from those of the initial UCFR design (Tak et al., 2013a) and this modification makes the burning speed of active core movement slower. It has been confirmed for both designs that a full-power operation of 60 years without refueling is feasible with respect to isotopics and criticality by a breed-and-burn strategy. The core performance characteristics of both designs have been evaluated in terms of axial/radial power shapes, neutron flux and nuclide distributions, breeding ratio, reactivity feedback coefficients, control rod worth, etc. By the design optimization study in this paper, the reductions of maximum neutron flux, fast neutron fluence, and axial/radial power peaking have been achieved, which are favorable for the safety of the UCFR.

Breed-and-burn↗

Convective Heat Transfer in Decay Heat Exchanger of Sodium-cooled Fast Reactor

Sodium-to-sodium heat exchangers such as the decay heat exchanger (DHX) are key components of the prototype generation [V sodium-cooled fast reactor (PGSFR). The DHX transfers the decay heat generated in the reactor core to the decay heat removal loop after a reactor shutdown. Shell-and-tube type with straight tube arrangement has been widely adopted for the sodium-to-sodium heat exchanger of the sodium-cooled fast reactor. Thus liquid metal heat transfer characteristics inside the tube and around the tube bundle have to be fully understood to design the sodium-to-sodium heat exchanger and evaluate its thermal performance accurately. Plenty of studies on the convective heat transfer of the liquid metals were conducted to present the Nusselt number (Nu) correlations with respect to the Peclet number (Pe). However, those correlations still deviated from each other considerably and few of correlations for low Pe range (less than 100) have been reported. Accordingly, it is difficult for a design code developer to choose a proper heat transfer coefficient correlation for the sodium-to-sodium heat exchanger particularly in a low flow rate range. Recently, a separate effect test for the DHX has been completed using the STELLA-1 facility at KAERI according to the STELLA program which aims at establishment of separate and integral test database (DB) for the design of the PGSFR. The heat transfer performance test of the DHX was successfully conducted under major operating conditions. The objective of this work is to suggest proper convective heat transfer correlations for tube- and shell-sides of the sodium-to-sodium heat exchanger based on the test DB obtained from STELLA-1. The Pe range of the analyzed experimental data is concentrated on less then 100. The adequacy of the suggested correlations is evaluated by applying to the DHX design code, SHXSA. Based on the DHX test DB of the STELLA-1, the proper convective heat transfer correlations for the tube- and shell-side of sodium-to-sodium heat exchangers were suggested. Two correlation combinations for each side were proposed and turned out to be reasonable through the evaluation using the SHXSA design code: Suggestion no. 1: - Lyon correlation (constant = 4) for tube-side; - Lubarsky-Kaufman correlation for shell-side. Suggestion no. 2: - Modified Lyon correlation (constant = 1.45) for tube-side; - Schad-modified correlation for shell-side The Pe ranges of the DHX test DB were 53 - 225 for the tube-side and 45 - 148 for the shell-side, but most of test conditions were concentrated on the Pe range of less than 100. Two correlation suggestions can be practically used for the design of sodium-to-sodium heat exchangers. However, by utilizing more test DB, it needs to be verified which one is closer to real phenomena. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Test Plan Proposal for an International Joint Project for Transient Testing of Fast Reactor Fuels

EXECUTIVE SUMMARY An international joint project for transient testing of fast reactor fuels is proposed. Input from several countries is provided regarding their advanced/fast reactor programs, and the objectives and methods are given based on both historical projects and modern tools and methods. The full package would include irradiation testing using TREAT, use of existing facilities for non destructive exams/evaluations (NDE) and post irradiation exams (PIE), and integration of current state-of-the-art modeling and simulation (M&S) tools to aid in both experiment design, and prediction of experimental results.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

CFD Evaluation of Pressure Change Along Coolant Passages in Sodium-Cooled Fast Reactor with Nek5000

To support the design efforts of advanced sodium-cooled fast reactors (SFRs), a series of computational fluid dynamics (CFD) simulations are performed to investigate the pressure change along various flow passages in the proposed SFR system. The simulations are carried out with the state-of-the-art spectral element flow solver, Nek5000. Two specific case studies are presented in this paper: the flow exiting the axial neutron reflector channels and the flow entering the fuel pin bundle. Due to the high Reynolds numbers expected, a Reynolds-averaged Navier-Stokes (RANS) approach is necessary to model the turbulence. A newly developed regularized k-omega RANS model is adopted in the related CFD calculations. The first case study explores the effect of Reynolds number on the pressure change when flow exits the reflector channels. The pressure change in this case has two major contributors: the change due to wall friction and the Bernoulli effect. It is noted that the nondimensional pressure loss follows a log-linear trend up to Re = 10 5 , and then the trend is flattened. In the second case study, the advanced NekNek coupling capability is tested where an integral domain can be divided into multiple subdomains with coupling interfaces, which would greatly ease the meshing process of complex engineering geometries and potentially save computational resources. The preliminary results obtained so far confirm the consistency between the NekNek results and those produced by regular Nek5000 simulation. The presented work demonstrates the readiness and flexibility of the related CFD techniques, which is part of the broader effort to leverage cutting-edge CFD to inform the advanced nuclear reactor designs.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Assessment of Startup Fuel Options for a Test or Demonstration Fast Reactor

This document explores startup fuel options for a proposed test/demonstration fast reactor. The fuel options considered are the metallic fuels U-Zr and U-Pu-Zr and the ceramic fuels UO 2 and UO 2 -PuO 2 (MOX). Attributes of the candidate fuel choices considered were feedstock availability, fabrication feasibility, rough order of magnitude cost and schedule, and the existing irradiation performance database. The reactor-grade plutonium bearing fuels (U-Pu-Zr and MOX) were eliminated from consideration as the initial startup fuels because the availability and isotopics of domestic plutonium feedstock is uncertain. There are international sources of reactor grade plutonium feedstock but isotopics and availability are also uncertain. Weapons grade plutonium is the only possible source of Pu feedstock in sufficient quantities needed to fuel a startup core. Currently, the available U.S. source of (excess) weapons-grade plutonium is designated for irradiation in commercial light water reactors (LWR) to a level that would preclude diversion. Weapons-grade plutonium also contains a significant concentration of gallium. Gallium presents a potential issue for both the fabrication of MOX fuel as well as possible performance issues for metallic fuel. Also, the construction of a fuel fabrication line for plutonium fuels, with or without a line to remove gallium, is expected to be considerably more expensive than for uranium fuels. In the case of U-Pu-Zr, a relatively small number of fuel pins have been irradiated to high burnup, and in no case has a full assembly been irradiated to high burnup without disassembly and re-constitution. For MOX fuel, the irradiation database from the Fast Flux Test Facility (FFTF) is extensive. If a significant source of either weapons-grade or reactor-grade Pu became available (i.e., from an international source), a startup core based on Pu could be reconsidered.

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗

Passive acoustic leak detection for sodium cooled fast reactors using hidden Markov models

Acoustic leak detection for steam generators of sodium fast reactors have been an active research topic since the early 1970's and several methods have been tested over the years. Inspired by its success in the field of automatic speech recognition, we here apply hidden Markov models (HMM) in combination with Gaussian mixture models (GMM) to the problem. To achieve this, we propose a new feature calculation scheme, based on the temporal evolution of the power spectral density (PSD) of the signal. Using acoustic signals recorded during steam/water injection experiments done at the Indira Gandhi Centre for Atomic Research (IGCAR), the proposed method is tested. We perform parametric studies on the HMM+GMM model size and demonstrate that the proposed method a) performs well without a priori knowledge of injection noise, b) can incorporate several noise models and c) has an output distribution that simplifies false alarm rate control. (authors)

21 SPECIFIC NUCLEAR REACTORS AND ASSOCIATED PLANTS↗