Nuclear Data Uncertainty - Impact on keff for the NBSR
Explore the source record for details and available documents.
Engineering topics
Publications and source records attributed to Cheng, Lap-Yan.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
Explore the source record for details and available documents.
The outline of the slides include: Motivations of the work; Modeling and simulation; Machine learning model; Results and comparison study with linear regression; and Conclusions. This work was done to help PBR designers and operators understand the burnup measurement better. We look forward to discussing the results in detail with industrial collaborators.
Pebble Bed Reactors are fueled with fuel pebbles that are circulated multiple times through the reactor vessel before discharge. During the normal operation of a PBR, ejected pebbles are returned to the reactor or discharged depending on the fuel burnup and physical condition of the pebbles. The burnup measurement is usually based on detected radiation signatures of fission products accumulated in the pebble fuel over burnup. Previous research has shown that height of photopeaks of fission products, such as 134Cs, 137Cs, 154Eu, etc., can be used independently or in combination to infer or predict the level of burnup in the fuel. However, it remains challenging to measure such complex sources due to self-shielding effects, strong radiation background and intervening materials. Another operational challenge is the required high throughput of burnup measurement, which necessitates limited measurement time and thus impacts quality of measured gamma-ray spectra. Hence, advanced spectral analysis methods are needed to analyze the noisy gamma spectra and predict the burnup values. We propose to use machine learning (ML) method to interpret gamma-ray spectra and predict the burnup values of the pebbles. ML has achieved widespread success and adoption across a few domains that require pattern recognition and analysis in varied data types. In this work, we apply three proven ML approaches - multilayer perceptrons, convolutional neural networks, and transformers - to the task of predicting fuel burnup from measured gamma spectra, and compile a dataset of simulated spectra for training and validation of the ML models. In this paper, we will discuss the network architecture of these three ML approaches and compare the performance of the simplest of these (MLP) to a standard linear regression.
Explore the source record for details and available documents.
Pebble Bed Reactors are fueled with fuel pebbles that are circulated multiple times through the reactor vessel before discharge. During the normal operation of a PBR, ejected pebbles are returned to the reactor or discharged depending on the fuel burnup and physical condition of the pebbles. The burnup measurement is usually based on detected radiation signatures of fission products accumulated in the pebble fuel over burnup. Previous research has shown that height of photopeaks of fission products, such as 134 Cs, 137 Cs, 154 Eu, etc., can be used independently or in combination to infer or predict the level of burnup in the fuel. However, it remains challenging to measure such complex sources due to self-shielding effects, strong radiation background and intervening materials. Another operational challenge is the required high throughput of burnup measurement, which necessitates limited measurement time and thus impacts quality of measured gamma-ray spectra. Hence, advanced spectral analysis methods are needed to analyze the noisy gamma spectra and predict the burnup values. We propose to use machine learning (ML) method to interpret gamma-ray spectra and predict the burnup values of the pebbles. ML has achieved widespread success and adoption across a few domains that require pattern recognition and analysis in varied data types. In this work, we apply three proven ML approaches - multilayer perceptrons, convolutional neural networks, and transformers - to the task of predicting fuel burnup from measured gamma spectra, and compile a dataset of simulated spectra for training and validation of the ML models. In this paper, we will discuss the network architecture of these three ML approaches and compare the performance of the simplest of these (MLP) to a standard linear regression.
Investigate the feasibility of machine learning in improving effectiveness and efficiency of safeguards at PBRs
Reduce regulatory uncertainty for prospective licensees of advanced reactors (ARs).
The Generation IV International Forum (GIF) formed a proliferation resistance and physical protection (PR&PP) Working Group (PRPPWG) to develop a methodology to evaluate the six GIF reactor technologies. The PRPPWG developed the methodology through a series of development and demonstration case studies and many insights were gained from the process. The PRPPWG, in collaboration with representatives of the GIF System Steering Committees/provisional Steering Committees (SSCs/pSSCs) for each of the six GIF reactor technologies, analyzed and evaluated the PR&PP aspects of the six reactor technologies. The outcome of this activity is documented in a joint PRPPWG and SSCs/pSSCs report made publicly available in 2011, highlighting the analysis performed and its main outcomes. Currently, the PRPPWG and the six GIF SSCs/pSSCs are in the process of repeating the exercise, based on the status of the six GIF system design concepts, considering the designs’ evolution since 2011. In line with the PR&PP by Design concept, whilst during the first exercise the level of detail of the design options enabled only a high-level evaluation of the systems’ proliferation resistance, the early design stage allowed designers to introduce modifications for improving the systems’ performance. With the designs’ current level of development, the opportunity for proposing modifications diminishes, whilst at the same time increasing the opportunity to inform the safeguards R&D community on potential R&D needs to meet safeguardability issues. The paper will first summarize the main concepts behind the PR&PP evaluation methodology. It will then present the main outcomes and lessons learned as in the current white papers on the PR aspects of the six GIF design concepts and will finally describe the ongoing activity to update them within a PR&PP-by-Design context.
The Generation Four International Forum (GIF) formed a proliferation resistance and physical protection (PR&PP) Working Group (PRPPWG) to develop a methodology to evaluate the six GIF reactor technologies. The PRPPWG developed the methodology through a series of development and demonstration case studies, using a hypothetical “Example Sodium Fast Reactor” (ESFR). The ESFR assessment enabled testing of the full methodology on a complete nuclear energy system and many insights were gained from the process. The PRPPWG, in collaboration with representatives of the GIF System Steering Committees/provisional Steering Committees (SSCs/pSSCs) for each of the six GIF reactor technologies, analyzed the PR&PP aspects of the six reactor technologies. The outcome of this activity is documented in a joint PRPPWG and SSCs/pSSCs report made publicly available in 2011, highlighting the analysis performed and its main outcomes. Currently, the PRPPWG and the six GIF SSCs/pSSCs are in the process of repeating the exercise, based on the current status of the six GIF system design concepts, taking into account the designs’ evolution since 2011. In line with Safeguards by Design concept, whilst during the first exercise the level of detail of the design options enabled only a high-level evaluation of the systems’ proliferation resistance, the early design stage allowed designers to introduce modifications for improving the systems’ performance. With the designs’ current level of detail, the opportunity for proposing modifications diminishes, whilst at the same time increasing the opportunity to inform the safeguards R&D community on potential R&D needs to meet safeguardability issues. The paper will first summarize the main concepts behind the PR&PP evaluation methodology. It will then summarize the main outcomes and lessons learned, as in the current white papers on the PR aspects of the six GIF design concepts, and will finally describe the ongoing activity to update them.
The Generation IV International Forum (GIF) “Technology Roadmap” identified proliferation resistance and physical protection (PR&PP) as one of the four goal areas to advance nuclear energy into its next, “fourth” generation. It recommended the development of a methodology to define measures for PR&PP and to evaluate them for the six nuclear energy systems selected by GIF. Accordingly, the Forum formed a PR&PP Working Group (PRPPWG) to develop a methodology and to define the parameters against which PR&PP would be measured. The PRPPWG developed the methodology through a series of development and demonstration case studies, by use of a hypothetical “Example Sodium Fast Reactor” (ESFR). The PR&PP ESFR assessment was the first opportunity to exercise the full methodology on a complete system, and many insights were gained from the process. The PRPPWG and representatives of the GIF System Steering Committees (SSCs) for each of the six GIF design concepts carried out a study on the PR&PP aspects of the six GIF designs. This interaction allowed the PRPPWG to engage the designers and to raise awareness about PR&PP. A number of international workshops have also been held which have introduced the methodology to design groups and other stakeholders. In this paper, we first summarize the technical progress and the major accomplishments of the PRPPWG: the evaluation methodology and the ESFR case study used to develop and test the methodology. The interaction with the GIF System Steering Committees and the relaunching of the activity is then presented. Finally, an outline of the future challenges is provided along the lines identified in the GIF R&D outlook 2018 report.
The U.S. Department of Energy (DOE) is preparing to re establish the capability to conduct transient testing of nuclear fuels at the Idaho National Laboratory (INL) Transient Reactor Test (TREAT) facility. The original TREAT core went critical in February 1959 and operated for more than 6,000 reactor startups before plant operations were suspended in 1994. DOE is now planning to restart the reactor using the plant's original high-enriched uranium (HEU) fuel. At the same time, the National Nuclear Security Administration (NNSA) Office of Material Management and Minimization Reactor Conversion Program is supporting analyses and fuel fabrication studies that will allow for reactor conversion to low-enriched uranium (LEU) fuel (i.e., fuel with less than 20% by weight 235U content) after plant restart. The TREAT Conversion Program's objectives are to perform the design work necessary to generate an LEU replacement core, to restore the capability to fabricate TREAT fuel element assemblies, and to implement the physical and operational changes required to convert the TREAT facility to use LEU fuel.
The research reactor (aka NBSR) at the National Institute of Standards and Technology (NIST) is a heavy water cooled, moderated, and reflected tank-type research reactor that operates at atmospheric pressure and a design power of 20 MW. Core power is generated in 30 fuel elements, each with 17 fuel plates in an upper section and a separate lower section. Power is removed using forced convection and heat exchangers in the cooling loop. The licensing analysis of the loss-of-coolant accident (LOCA) in the NBSR was addressed in the current Safety Analysis Report (SAR). The SAR discusses the low probability of a significant pipe break: 'the main piping is located in protected areas, system pressures are low, and flow rates are small so that wear is not an issue.' Therefore, the probability of a large break, including a double-ended guillotine break, is extremely low. For smaller breaks where the operator has time to take action, procedures are in place to 1) mitigate a loss of water by tripping pumps and closing control valves to isolate the reactor vessel after the falling water level in the vessel is detected by instrumentation, and 2) ensure that emergency cooling water continues to flow for as long as needed. The SAR also refers to an analysis of why there is sufficient water without forced cooling to cool the fuel elements (FEs) in a LOCA if the flow channels between two fuel plates (or a fuel plate and an outside plate) remain filled with water. That analysis was based on a flooding-limited critical power but assumed that reactor trip had taken place. A more recent study showed that some scenarios may lead to the draining of water from the coolant channels. In these cases, the cooling on the inside of the fuel element is from emergency cooling water that falls only along the inner surface of one of the side plates in the fuel elements; the remaining surfaces within the fuel element being exposed to gas. The analysis therein was done for both the present high-enriched uranium fuel (HEU) and the proposed low-enriched uranium fuel (LEU). The LOCA analyses done previously assumed that there is sufficient cooling during the short-term, i.e., before the water level reaches the top of the fuel plates and begins to drain from the coolant channels or before decay heat becomes low enough such that the flooding-limited critical power shows that there is no problem. The present study is to determine if that assumption is correct, i.e., to consider the situation immediately after the break, before and after reactor trip when the power is still relatively high.
A study of the fuel temperature during the early phase of a loss-of-coolant accident (LOCA) in the NIST research reactor (NBSR) was completed. Previous studies had been reported in the preliminary safety analysis report for the conversion of the NBSR from high-enriched uranium (HEU) fuel to low-enriched (LEU) fuel. Those studies had focused on the most vulnerable LOCA situation, namely, a double-ended guillotine break in the time period after reactor trip when water is drained from either the coolant channels inside the fuel elements or the region outside the fuel elements. The current study fills in a gap in the analysis which is the early phase of the event when there may still be water present but the reactor is at power or immediately after reactor trip and pumps have tripped. The calculations were done, for both the current HEU-fueled core and the proposed LEU core, with the TRACE thermal-hydraulic systems code. Several break locations and different break sizes were considered. In all cases the increase in the clad (or fuel meat) temperature was relatively small so that a large margin to the temperature threshold for blistering (the Safety Limit for the NBSR) remained.
The framework for Proliferation Resistance and Physical Protection (PR & PP) evaluation is to define a set of challenges, to obtain the system responses, and to assess the outcomes. The assessment of outcomes heavily relies on pathways, defined as sequences of events or actions that could potentially be followed by a State or a group of individuals in order to achieve a proliferation objective, with the defined threats as initiating events. There may be large number of segments connecting pathway stages (e.g. acquisition, processing, and fabrication for PR) which can lead to even larger number of pathways or scenarios through possible different combinations of segment connections, each with associated probabilities contributing to the overall risk. Clustering of these scenarios in specified stage attribute intervals is important for their tractable analysis and outcome assessment. A software tool for scenario generation and clustering (OSUPR) is developed that utilizes the PRCALC code developed at the Brookhaven National Laboratory for scenario generation and the K- means, mean shift and adaptive mean shift algorithms as possible clustering schemes. The results of the study using the Example Sodium Fast Breeder as an example system show that clustering facilitates the probabilistic or deterministic analysis of scenarios to identify system vulnerabilities and communication of the major risk contributors to stakeholders. The results of the study also show that the mean shift algorithm has the most potential for assisting the analysis of the scenarios generated by PRCALC.
The safety goal of the current designs of advanced high‐temperature thermal gas‐cooled reactors (HTRs) is that no core meltdown would occur in a depressurization event with a combination of concurrent safety system failures. This study focused on the analysis of passive decay heat removal (DHR) in a GEN IV direct‐cycle gas‐cooled fast reactor (GFR) which is based on the technology developments of the HTRs. Given the different criteria and design characteristics of the GFR, an approach different from that taken for the HTRs for passive DHR would have to be explored. Different design options based on maintaining core flow were evaluated by performing transient analysis of a depressurization accident using the system code RELAP5‐3D. The study also reviewed the conceptual design of autonomous systems for shutdown decay heat removal and recommends that future work in this area should be focused on the potential for Brayton cycle DHRs.