Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “SAFETY FACTOR”

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 253 records · Page 14

Comparison of Edge Plasma Perturbation During ELM Control Using One vs Two Toroidal Rows of RMP Coils in ITER Similar Shaped Plasmas on DIII-D

Large Type-I edge localized modes (ELMs) were suppressed by n = 3 resonant magnetic perturbations (RMPs) from a set of internal coils (I-coil) in plasmas with an ITER similar shape at the ITER pedestal collisionality, {nu}*{sub e} {approx} 0.1 and low edge safety factor (q{sub 95} {approx} 3.6), with either a single toroidal row of the internal RMP coils or two poloidally separated rows of coils. ELM suppression with a single row of internal coils was achieved at approximately the same q{sub 95} surface-averaged perturbation field as with two rows of coils, but required higher current per coil. Maintaining complete suppression of ELMs using n = 3 RMPs from a single toroidal row of internal coils was less robust to variations in input neutral beam injection torque than previous ELM suppression cases using both rows of internal coils. With either configuration of RMP coils, maximum ELM size is correlated with the width of the edge region having good overlap of the magnetic islands from vacuum field calculations.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Assessment of Alaska's North Slope Oil Field Capacity to Sequester CO{sub 2}

The capacity of 21 major fields containing more than 95% of the North Slope of Alaska's oil were investigated for CO{sub 2} storage by injecting CO{sub 2} as an enhanced oil recovery (EOR) agent. These fields meet the criteria for the application of miscible and immiscible CO{sub 2}-EOR methods and contain about 40 billion barrels of oil after primary and secondary recovery. Volumetric calculations from this study indicate that these fields have a static storage capacity of 3 billion metric tons of CO{sub 2}, assuming 100% oil recovery, re-pressurizing the fields to pre-fracturing pressure and applying a 50% capacity reduction to compensate for heterogeneity and for water invasion from the underlying aquifer. A ranking produced from this study, mainly controlled by field size and fracture gradient, identifies Prudhoe, Kuparuk, and West Sak as possessing the largest storage capacities under a 20% safety factor on pressures applied during storage to avoid over-pressurization, fracturing, and gas leakage. Simulation studies were conducted using CO{sub 2} Prophet to determine the amount of oil technically recoverable and CO{sub 2} gas storage possible during this process. Fields were categorized as miscible, partially miscible, and immiscible based on the miscibility of CO{sub 2} with their oil. Seven sample fields were selected across these categories for simulation studies comparing pure CO{sub 2} and water-alternating-gas injection. Results showed that the top two fields in each category for recovery and CO{sub 2} storage were Alpine and Point McIntyre (miscible), Prudhoe and Kuparuk (partially miscible), and West Sak and Lisburne (immiscible). The study concludes that 5 billion metric tons of CO{sub 2} can be stored while recovering 14.2 billion barrels of the remaining oil.

02 PETROLEUM↗

Numerical Analysis of the Effects of Normalized Plasma Pressure on RMP ELM Suppression in DIII-D

The effect of normalized plasma pressure as characterized by normalized pressure parameter (beta(N)) on the suppression of edge localized modes (ELMs) using resonant magnetic perturbations (RMPs) is studied in low-collisionality (nu* <= 0.2) H-mode plasmas with low-triangularity ( = 0.25) and ITER similar shapes ( = 0.51). Experimental results have suggested that ELM suppression by RMPs requires a minimum threshold in plasma pressure as characterized by beta(N). The variations in the vacuum field topology with beta(N) due to safety factor profile and island overlap changes caused by variation of the Shafranov shift and pedestal bootstrap current are examined numerically with the field line integration code TRIP3D. The results show very small differences in the vacuum field structure in terms of the Chirikov (magnetic island overlap) parameter, Poincare sections and field line loss fractions. These differences do not appear to explain the observed threshold in beta(N) for ELM suppression. Linear peeling-ballooning stability analysis with the ELITE code suggests that the ELMs which persist during the RMPs when beta(N) is below the observed threshold are not type I ELMs, because the pedestal conditions are deep within the stable regime for peeling-ballooning modes. These ELMs have similarities to type III ELMs or low density ELMs.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Relation between energetic and standard geodesic acoustic modes

Geodesic Acoustic Modes (GAMs) are electrostatic, axisymmetric modes which are non-linearly excited by turbulence. They can also be excited linearly by fast-particles; they are then called Energetic-particle-driven GAMs (EGAMs). Do GAMs and EGAMs belong to the same mode branch? Through a linear, analytical model, in which the fast particles are represented by a Maxwellian bump-on-tail distribution function, we find that the answer depends on several parameters. For low values of the safety factor q and for high values of the fast ion energy, the EGAM originates from the GAM. On the contrary, for high values of q and for low values of the fast ion energy, the GAM is not the mode which becomes unstable when fast particles are added: the EGAM then originates from a distinct mode, which is strongly damped in the absence of fast particles. The impact of other parameters is further explored: ratio of the ion temperature to the electron temperature, width of the fast particle distribution, mass and charge of the fast ions. The ratio between the EGAM and the GAM frequencies was found in experiments (DIII-D) and in non-linear numerical simulations (code GYSELA) to be close to 1/2: the present analytical study allows one to recover this ratio.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Alfv?nic Instabilities and Fast Ion Transport in the DIII-D Tokamak

Neutral beam injection into reversed magnetic shear DIII-D plasmas produces a variety of Alfvenic activity including Toroidicity and Ellipticity induced Alfven Eigenmodes (TAE/EAE, respectively) and Reversed Shear Alfven Eigenmodes (RSAE) as well as their spatial coupling. These modes are typically studied during the discharge current ramp phase when incomplete current penetration results in a high central safety factor and strong drive due to multiple higher order resonances. During this same time period Fast-Ion D{sub {alpha}} (FIDA) spectroscopy shows that the central fast ion profile is flattened, the degree of which depends on the Alfven eigenmode amplitude. Interestingly, localized electron cyclotron heating (ECH) near the mode location stabilizes RSAE activity and results in significantly improved fast ion confinement relative to discharges with ECH deposition on axis. In these discharges, RSAE activity is suppressed when ECH is deposited near the radius of the shear reversal point and enhanced with deposition near the axis. To simulate the observed neutral beam ion redistribution, NOVA calculations of the 3D eigenmode structures are matched with experimental measurements and used in combination with the ORBIT guiding center following code. For fixed frequency eigenmodes, it is found that ORBIT calculations cannot explain the observed beam ion transport with experimentally measured mode amplitudes. Possible explanations are considered including recent simulation results incorporating eigenmodes with time dependent frequencies.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Limit Load and Buckling Analysis for Assessing Hanford Single-Shell Tank Dome Structural Integrity - 12278

The U.S. Department of Energy, Office of River Protection has commissioned a structural analysis of record for the Hanford single shell tanks to assess their structural integrity. The analysis used finite element techniques to predict the tank response to the historical thermal and operating loads. The analysis also addressed the potential tank response to a postulated design basis earthquake. The combined response to static and seismic loads was then evaluated against the design requirements of American Concrete Institute standard, ACI-349-06, for nuclear safety-related concrete structures. Further analysis was conducted to estimate the plastic limit load and the elastic-plastic buckling capacity of the tanks. The limit load and buckling analyses estimate the margin between the applied loads and the limiting load capacities of the tank structure. The potential for additional dome loads from waste retrieval equipment and the addition of large dome penetrations to accommodate retrieval equipment has generated additional interest in the limit load and buckling analyses. This paper summarizes the structural analysis methods that were used to evaluate the limit load and buckling of the single shell tanks. This paper summarizes the structural analysis methods that were used to evaluate the limit load and buckling limit states of the underground single shell tanks at the Hanford site. The limit loads were calculated using nonlinear finite element models that capture the progressive deformation and damage to the concrete as it approaches the limit load. Both uniform and concentrated loads over the tank dome were considered, and the analysis shows how adding a penetration in the center of the tank would affect the limit loads. For uniform surface loads, the penetration does not affect the limit load because concrete crushing and rebar yielding initiates first at the top of the wall, away from the penetration. For concentrated loads, crushing initiates at the center of the dome, so the penetration does reduce the concentrated limit load somewhat. However, the safety factors comparing the limit loads to the maximum allowable applied loads remain well above the required value of 3.0. The buckling analysis method accounts for the geometric imperfections, concrete creep, cracking and reinforcements, and concrete plasticity in determining the allowable buckling load limits. The method was demonstrated in this paper for the evaluation of a tank before and after a penetration is added in the dome center. Finite element buckling models were used to accurately calculate the linear critical buckling loads. The models showed that adding the penetration reduces the linear critical buckling load by only 2.5%. Bounding cases also showed that the possible range of soil support on the walls does not significantly affect the dome buckling loads. Buckling models of the full 360 deg. dome also showed that the dome is more resistant to buckling when the load is offset than when it is positioned over the center. These limit load and buckling analysis methods are being used at the Hanford site to assess the tank loads that can be safely applied during future waste retrieval activities. (authors)

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

The Adaptable and Resilient Safety System: The Human Factor in Future In-Time Aviation Safety Management Systems

In-time integrated safety management will be paramount for safely enabling the envisioned transformations of the future National Airspace System (NAS). The path for realizing the vision includes addressing the increasing need for advanced data analytics and fusion of aviation safety data, managed by human decision-makers. The paper describes safety management systems and its’ challenges, and how the concept of In-time Aviation Safety Management Systems addresses the need to ensure an adaptable and resilient future safety system in the envisioned transformed NAS. Finally, it discusses potential human factors challenges, including new human roles and responsibilities, new information and cognitive requirements, new intelligent technologies that change human-system interaction and coordination, and new design paradigms for human system integration and teaming.

L. Prinzel↗

Fully Nonlinear Edge Gyrokinetic Simulations of Kinetic Geodesic-Acoustic Modes and Boundary Flows

We present edge gyrokinetic neoclassical simulations of tokamak plasmas using the fully nonlinear (full-f) continuum code TEMPEST. A nonlinear Boltzmann model is used for the electrons. The electric field is obtained by solving the 2D gyrokinetic Poisson Equation. We demonstrate the following: (1) High harmonic resonances (n > 2) significantly enhance geodesic-acoustic mode (GAM) damping at high-q (tokamak safety factor), and are necessary to explain both the damping observed in our TEMPEST q-scans and experimental measurements of the scaling of the GAM amplitude with edge q{sub 95} in the absence of obvious evidence that there is a strong q dependence of the turbulent drive and damping of the GAM. (2) The kinetic GAM exists in the edge for steep density and temperature gradients in the form of outgoing waves, its radial scale is set by the ion temperature profile, and ion temperature inhomogeneity is necessary for GAM radial propagation. (3) The development of the neoclassical electric field evolves through different phases of relaxation, including GAMs, their radial propagation, and their long-time collisional decay. (4) Natural consequences of orbits in the pedestal and scrape-off layer region in divertor geometry are substantial non-Maxwellian ion distributions and flow characteristics qualitatively like those observed in experiments.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Kink modes and surface currents associated with vertical displacement events

The fast termination phase of a vertical displacement event (VDE) in a tokamak is modeled as a sequence of shrinking equilibria, where the core current profile remains constant so that the safety-factor at the axis, q{sub axis}, remains fixed and the q{sub edge} systematically decreases. At some point, the n = 1 kink mode is destabilized. Kink modes distort the magnetic field lines outside the plasma, and surface currents are required to nullify the normal component of the B-field at the plasma boundary and maintain equilibrium at finite pressure. If the plasma touches a conductor, the current can be transferred to the conductor, and may be measurable by the halo current monitors. This report describes a practical method to model the plasma as it evolves during a VDE, and determine the surface currents, needed to maintain equilibrium. The main results are that the onset conditions for the disruption are that the growth-rate of the n = 1 kink exceeds half the Alfven time and the associated surface current needed to maintain equilibrium exceeds one half of the core plasma current. This occurs when q{sub edge} drops below a low integer, usually 2. Application to NSTX provides favorable comparison with non-axisymmetric halo-current measurements. The model is also applied to ITER and shows that the 2/1 mode is projected to be the most likely cause of the final disruption.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Operating Characteristics in DIII-D ELM-Suppressed RMP H-modes with ITER Similar Shapes

Fast energy transients, incident on the DIII-D divertors due to Type-I edge localized modes (ELMs), are eliminated using small dc currents in a simple set of non-axisymmetric coils that produce edge resonant magnetic perturbations (RMP). In ITER similar shaped (ISS) plasmas, with electron pedestal collisionalities matched to those expected in ITER a sharp resonant window in the safety factor at the 95 percent normalized poloidal flux surface is observed for ELM suppression at q{sub 95}=3.57 with a minimum width {delta}q{sub 95} of {+-}0.05. The size of this resonant window has been increased by a factor of 4 in ISS plasmas by increasing the magnitude of the current in an n=3 coil set along with the current in a separate n=1 coil set. The resonant ELM-suppression window is highly reproducible for a given plasma shape, coil configuration and coil current but can vary with other operating conditions such as {beta}{sub N}. Isolated resonant windows have also been found at other q95 values when using different RMP coil configurations. For example, when the I-coil is operated in an n=3 up-down asymmetric configuration rather than an up-down symmetric configuration a resonant window is found near q{sub 95}=7.4. A Fourier analysis of the applied vacuum magnetic field demonstrates a statistical correlation between the Chirikov island overlap parameter and ELM suppression. These results have been used as a guide for RMP coil design studies in various ITER operating scenarios.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Non-inductive current drive and transport in high beta(N) plasmas in JET

A route to stationary MHD stable operation at high beta(N) has been explored at the Joint European Torus (JET) by optimizing the current ramp-up, heating start time and the waveform of neutral beam injection (NBI) power. In these scenarios the current ramp-up has been accompanied by plasma pre-heat (or the NBI has been started before the current flat-top) and NBI power up to 22 MW has been applied during the current flat-top. In the discharges considered transient total beta(N) approximate to 3.3 and stationary (during high power phase) beta(N) approximate to 3 have been achieved by applying the feedback control of beta(N) with the NBI power in configurations with monotonic or flat core safety factor profile and without an internal transport barrier (ITB). The transport and current drive in this scenario is analysed here by using the TRANSP and ASTRA codes. The interpretative analysis performed with TRANSP shows that 50-70% of current is driven non-inductively; half of this current is due to the bootstrap current which has a broad profile since an ITB was deliberately avoided. The GLF23 transport model predicts the temperature profiles within a +/- 22% discrepancy with the measurements over the explored parameter space. Predictive simulations with this model show that the E x B rotational shear plays an important role for thermal ion transport in this scenario, producing up to a 40% increase of the ion temperature. By applying transport and current drive models validated in self-consistent simulations of given reference scenarios in a wider parameter space, the requirements for fully non-inductive stationary operation at JET are estimated. It is shown that the strong stiffness of the temperature profiles predicted by the GLF23 model restricts the bootstrap current at larger heating power. In this situation full non-inductive operation without an ITB can be rather expensive strongly relying on the external non-inductive current drive sources.

99 GENERAL AND MISCELLANEOUS↗

Requirements and Conceptual Design of Off-gas Systems for the Reprocessing of Metallic Fuels

An assessment has been conducted to determine how key regulations regarding volatile radionuclide emissions to the atmosphere may apply to the off-gas streams associated with electrochemical reprocessing. The scope of this assessment was based upon a generic electrochemical reprocessing scheme with a throughput rate of 200 MTIHM/y applied to metallic fuel discharged from a sodium fast reactor (SFR), but the findings are able to be translated to other advanced nuclear scenarios as merited. Air dispersion modeling was performed using the EPA CAP-88 model and evaluated the uncontrolled decontamination factors (DFs) that would be required to achieve regulatory compliance with the dose-based limits set forth by EPA regulation 40 CFR 190.10(a). These DFs were compared to those required by fuel cycle–based limits set forth by EPA regulation 40 CFR 190.10(b). Two theoretical sites with disparate climatological conditions were selected for air dispersion modeling (Idaho and Tennessee). The radionuclides modeled included 3 H, 85 Kr, 129 I, and selected alpha-emitting transuranic isotopes (referred to here as 239 Pu-TRU <1y ). It was found that the fuel cycle-based limits in 40 CFR 190.10(b) are most restrictive for 85 Kr and 239 Pu-TRU <1y , with DFs of 3 and 6.1E+09, respectively. The dose-based limit as derived from 40 CFR 190.10(a) could require mitigation of tritium in some scenarios, with an estimated DF of about 3 for the reference scenarios. The fuel cycle-based limit for 129 I resulted in a DF of about 240 for the reference scenario. The need for iodine mitigation based on dose to the public depended upon the physical form of iodine as either particulate or vapor-phase species. Emission of iodine from the facility as a vapor necessitated DFs of about 2 but emission as a particulate would require DFs >6,000 to meet thyroid dose-based limits. Effects of physical form on needed iodine mitigation are significant, but the understanding of speciation of iodine both during electrochemical reprocessing and after release to the atmosphere is limited. The electrochemical processing unit operations were evaluated to identify potential release points for the volatile radionuclides and to assess the potential for retention of the radionuclides within the process (thus decreasing the need for mitigation). Mitigation strategies for 3 H, 85 Kr, 129 I, and 239 Pu-TRU <1y were identified. In all cases, there are reasonably achievable pathways to regulatory compliance, although in some cases additional R&D is merited to verify the chemical speciation of these isotopes and to develop and demonstrate potential treatment technologies for this application. Whether or not additional off-gas controls (beyond common operations such as HEPA filtration and oxygen and moisture control) are needed for any of these regulated or volatile radionuclides depends on the (a) type of facility (NRC-regulated or DOE), (b) used fuel process rate, (c) used fuel burnup and composition, (d) speciation and retention of volatile radionuclides in the process and in the cell gas cleanup system, (e) site-specific parameters such as location, meteorology, stack height, and site boundaries, and (f) levels of conservatism and safety factors used in assessing compliance to air emissions regulations. Performance of this assessment revealed several areas where information is lacking or additional research is required in order to better determine if or what kinds of off-gas control might be needed. First, and most significantly, the understanding of the chemical speciation and physical form and partitioning of iodine during electrochemical processing operations is lacking and prevents the ability to accurately assess the potential iodine mitigation requirements. Future research in this area should be multifaceted and include thermodynamic modeling of iodine speciation in different process steps, experiments to quantify the kinetics of vapor-phase and melt-phase transitions, bench-scale experiments to determine the potential chemical and physical form of iodine emissions from the electrorefining process, and verification of iodine behavior with experiments utilizing operational facilities. Similarly, an improved understanding of iodine behavior in the environment after release from the facility stack will be required to refine dose estimations, as particulate and vapor-phase emissions result in significantly different doses to the MEI.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗

The Influence of Three-dimensional Stochastic Magnetic Boundaries on Plasma Edge Transport and the Resulting Plasma Wall Interaction

The three-dimensional (3D) features of plasma edge profiles and wall interaction patterns induced by edge resonant magnetic perturbation fields (RMP) are discussed comparing TEXTOR and DIII-D. We show that the scrape-off layer (SOL) profiles and decay lengths depend the edge safety factor, the RMP base mode as well as on the plasma rotation during RMP application indicating modification of SOL transport by the 3D perturbation fields. This is compatible with channeling of particle and heat efflux along open perturbed field lines in the very edge of the plasma boundary into a completely re-arranged, helically striated 3D divertor footprint. The distribution of the measured divertor heat and particle fluxes at DIII-D match the vacuum modeled magnetic footprint topology in L-mode while in H-mode the striation width exceeds the modeled footprint width by 15-30%. This 3D structure of the measured heat and particle fluxes results in a new situation for the material erosion properties and initial quantification of the net-erosion within the 3D footprint shows in L-mode a 50% decrease of the chemical erosion yield and evidence for a comparably small 15-20% increase in physical sputtering. Extrapolation of these findings to ITER by vacuum modeling of the magnetic footprint for the actual ELM control coils shows a similar vacuum magnetic footprint topology as found at DIII-D during RMP ELM suppression. However, the open field lines escape the CFC covered ITER divertor area potentially transferring net-erosion characteristics from the CFC domain onto the Tungsten including so far unconsidered heat and particle loads on this sensitive material.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

A Study on the Use of XCT and FEA to Predict the Elastic Behavior of Additively Manufactured Parts of Cylindrical Geometry

Defining general criteria for the acceptability of defects within industrial components is often complicated, since the specific load conditions and the criticality of the given application should be considered individually. In order to minimize the risk of failure, high safety factors are commonly adopted during quality control. However this practice is likely to cause the rejection of components whose defects would be instead acceptable if a more sound knowledge of the component behaviour were achieved. Parts produced by additive manufacturing (AM) may suffer from various defects, including micro- or macro-holes, delamination and microstructural discontinuities. Such processes, which are specially suitable for one-off components, require robust and reliable inspection before a part is accepted or rejected, since the refusal of even a single part at the end of the production process represents a significant loss. For this reason, it would be very useful to simulate in a reliable way whether a certain defect is truly detrimental to the proper working of the part during operation or whether the component can still be used, despite the presence of a defect. To this purpose, the paper highlights the benefits of a synergistic interaction between Industrial X-ray computed tomography (XCT) and finite element analysis (FEA). Internal defects of additively manufactured parts can be identified in a non-destructive way by means of XCT. Then FEA can be performed on the XCT-based virtual model of the real component, rather than on the ideal CAD geometry. A proof of concept of this approach is proposed here for a reference construct produced in an Aluminium alloy by AM. Numerical results of the proposed combined XCT–FEA procedure are contrasted with experimental data from tensile tests. The findings sustain the reliability of the method and allow to assess its full provisional accuracy for parts of cylindrical geometry designed to operate in the elastic field. The paper moves a step beyond the present application limits of tomography as it is currently employed for AM parts and it evidences instead the possibility of extending the usage of tomography to acceptance testing and prediction of operative behaviour.

42 ENGINEERING↗

Estimating List-Mode Data Sensitivities to Nuclear Data with MCNP6

Nuclear data are a vital component of predictive simulations used in applications like experiment design, stockpile stewardship, nuclear nonproliferation/safeguards, health physics, and criticality safety. A singular simulation requires the coalescence of different areas of nuclear data such as cross sections, angular distributions, and energy distributions of emitted neutrons for different materials and energy ranges. Improving nuclear data and thus reducing the uncertainty in simulated parameters could enable smaller, better-informed safety factors and ultimately reduce operational and procedural costs. There is a constant effort to garner a better understanding of the physical quantities represented by nuclear data through experiments. Integral experiment benchmarks use simulated and measured results to validate current nuclear data values. In the past, benchmarks primarily focused on the effective multiplication factor (k eff ); however, this limited scope has caused compensating errors and areas of nuclear data that lack validation. Compensating errors are inaccuracies in nuclear data that are obfuscated by cancellation when observing integrated values such as k eff . Diverse integral benchmark experiments that look for quantities of interest other than k eff and include multiple responses minimize the possibility of compensating errors and provides validation to areas of nuclear data previously lacking experimental validation. Benchmark experiments can be optimized during the design process to be highly dependent on specific areas of nuclear data. The dependence of a response in an experiment to a specific area/type of nuclear data is defined as sensitivity. A larger sensitivity means that nuclear data uncertainties will play a larger role in the response(s) resulting in larger bias. Currently, the sensitivity capabilities of the Monte Carlo N-Particle (MCNP ®1 ) transport code are limited to responses of k eff and tallied values (e.g., flux, surface current). As a part of the EUCLID project, this work explores estimating list-mode nuclear data sensitivities that can be used to design experiments aimed to constrain and reduce compensating errors in nuclear data by focusing on responses other than k eff . Tallied values are ideal quantities that are estimated with detectors during experiments. List-mode data (a list of neutron collection times) are the direct output of detector systems in subcritical neutron noise experiments. Expanding MCNP sensitivity capabilities to include the sensitivity of responses estimated from list-mode data, such as the prompt neutron decay constant (α) and multiplicity estimates (S and D), enables more direct comparison of simulated and measured experimental quantities. Additionally, deterministic tools such as SENSMG are capable of obtaining sensitivities to a wide variety of responses; however, these tools cannot handle complex geometries due to the assumptions made in discretizing the phase-space variables of the Boltzman transport equation.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Implementation of AI/DEEP learning disruption predictor into a plasma control system

Abstract This paper reports on advances in the state‐of‐the‐art deep learning disruption prediction models based on the Fusion Recurrent Neural Network (FRNN) originally introduced in a 2019 NATURE publication [ https://doi.org/10.1038/s41586‐019‐1116‐4 ]. In particular, the predictor now features not only the “disruption score,” as an indicator of the probability of an imminent disruption, but also a “sensitivity score” in real time to indicate the underlying reasons for the imminent disruption. This adds valuable physics interpretability for the deep learning model and can provide helpful guidance for control actuators now implemented into a modern plasma control system (PCS). The advance is a significant step forward in moving from modern deep learning disruption prediction to real‐time control and brings novel AI‐enabled capabilities relevant for application to the future burning plasma ITER system. Our analyses use large amounts of data from JET and DIII‐D vetted in the earlier NATURE publication. In addition to “when” a shot is predicted to disrupt, this paper addresses reasons “why” by carrying out sensitivity studies. FRNN is accordingly extended to use more channels of information, including measured DIII‐D signals such as (i) the “n1rms” signal that is correlated with the n = 1 modes with finite frequency, including neoclassical tearing mode and sawtooth dynamics; (ii) the bolometer data indicative of plasma impurity control; and (iii) “q‐min”—the minimum value of the safety factor relevant to the key physics of kink modes. The additional channels and interpretability features expand the ability of the deep learning FRNN software to provide information about disruption subcategories as well as more precise and direct guidance for the actuators in a PCS.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

Investigation of Structure and Crack Formation in Welded Joints of Single Crystal Ni-Base Alloys

Conditions of initiation of hot cracks in welding of single crystal heat-resistant alloys with a 56% content of the γ’-phase were studied. Conditions for providing a polycrystalline structure in HAZ and part of the weld metal were artificially created in one region of a welded joint with the single crystal structure. Welding of the single crystal alloy was performed by the electron beam method in vacuum. Cracks in a low-temperature ductility dip range (ductility dip cracking) were found to form in HAZ and weld of a welded joint with the polycrystalline structure. The nature of cracking of the single crystal weld is related to formation of low-melting point eutectoid interlayers, as well as to exhausting of the safety factor for ductility. In conclusion, cracks in the single crystal weld along the strain grain boundaries are formed by a similar mechanism.

36 MATERIALS SCIENCE↗

Is Hyalella azteca a Suitable Model Leaf-Shredding Benthic Crustacean for Testing the Toxicity of Sediment-Associated Metals in Europe?

The leaf-shredding crustacean Hyalella azteca, which is indigenous to Northern and Central America, is used to assess environmental risks associated with (metal-)contaminated sediments and to propose sediment quality standards also in Europe. Yet, it is unknown if H. azteca is protective for European crustacean shredders. We thus compared the sensitivity of H. azteca with that of the European species Asellus aquaticus and Gammarus fossarum towards copper- and cadmium-contaminated sediments (prepared according to OECD 218) under laboratory conditions employing mortality and leaf consumption as endpoints. H. azteca either reacted approximately fourfold more sensitive than the most tolerant tested species (as for cadmium) or its sensitivity was only 1.6 times lower than the highest sensitivity determined (as for copper), which should be covered by safety factors applied during risk assessments. Therefore, the results for the sediment type and the two heavy metals tested during the present study in combination with the existence of standardized testing protocols, their ease of culture, and short generation time, suggest H. azteca as suitable crustacean model shredder for assessing the toxicity of sediment-associated metals in Europe.

54 ENVIRONMENTAL SCIENCES↗