Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Error Mitigation”

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 613 records · Page 34

FREQUENCY-DEPENDENT DISPERSION MEASURES AND IMPLICATIONS FOR PULSAR TIMING

The dispersion measure (DM), the column density of free electrons to a pulsar, is shown to be frequency dependent because of multipath scattering from small-scale electron-density fluctuations. DMs vary between propagation paths whose transverse extent varies strongly with frequency, yielding arrival times that deviate from the high-frequency scaling expected for a cold, uniform, unmagnetized plasma (1/frequency{sup 2}). Scaling laws for thin phase screens are verified with simulations; extended media are also analyzed. The rms DM difference across an octave band near 1.5 GHz is ∼ 4 × 10{sup −5} pc cm{sup −3} for pulsars at ∼1 kpc distance. The corresponding arrival-time variations are a few to hundreds of nanoseconds for DM ≲ 30 pc cm{sup −3} but increase rapidly to microseconds or more for larger DMs and wider frequency ranges. Chromatic DMs introduce correlated noise into timing residuals with a power spectrum of “low pass” form. The correlation time is roughly the geometric mean of the refraction times for the highest and lowest radio frequencies used, ranging from days to years, depending on the pulsar. We discuss implications for methodologies that use large frequency separations or wide bandwidth receivers for timing measurements. Chromatic DMs are partially mitigable by including an additional chromatic term in arrival time models. Without mitigation, an additional term in the noise model for pulsar timing is implied. In combination with measurement errors from radiometer noise, an arbitrarily large increase in total frequency range (or bandwidth) will yield diminishing benefits and may be detrimental to overall timing precision.

79 ASTRONOMY AND ASTROPHYSICS↗

HRA Aerospace Challenges

Compared to equipment designed to perform the same function over and over, humans are just not as reliable. Computers and machines perform the same action in the same way repeatedly getting the same result, unless equipment fails or a human interferes. Humans who are supposed to perform the same actions repeatedly often perform them incorrectly due to a variety of issues including: stress, fatigue, illness, lack of training, distraction, acting at the wrong time, not acting when they should, not following procedures, misinterpreting information or inattention to detail. Why not use robots and automatic controls exclusively if human error is so common? In an emergency or off normal situation that the computer, robotic element, or automatic control system is not designed to respond to, the result is failure unless a human can intervene. The human in the loop may be more likely to cause an error, but is also more likely to catch the error and correct it. When it comes to unexpected situations, or performing multiple tasks outside the defined mission parameters, humans are the only viable alternative. Human Reliability Assessments (HRA) identifies ways to improve human performance and reliability and can lead to improvements in systems designed to interact with humans. Understanding the context of the situation that can lead to human errors, which include taking the wrong action, no action or making bad decisions provides additional information to mitigate risks. With improved human reliability comes reduced risk for the overall operation or project.

DeMott, Diana↗

Template engine applied to rapid modeling

Modeling complex problems can require a significant effort, be subject to human error, and contain repeated patterns of input. SCALE 6.2 and the Used Nuclear Fuel-Storage, Transportation and Disposal Analysis Resource and Data System (UNF-ST-DARDS) have attempted to mitigate these problems with the use of a template engine and template syntax. The template syntax allows redundant input to be consolidated into its simplest pattern and parameterized on the parts that change, reducing effort and likelihood of errors and encapsulating user experience into a reusable model component. This paper presents a summary of rapid modeling benefits and techniques that are facilitated when using templates. Specifically, this paper will discuss how these are applied in conducting as-loaded spent nuclear fuel canister criticality calculations using UNF-ST-DARDS and model perturbation and sampling using SCALE 6.2. (authors)

42 ENGINEERING↗

ATTNChecker: Highly-Optimized Fault Tolerant Attention for Large Language Model Training

Large Language Models (LLMs) have demonstrated remarkable performance in various natural language processing tasks. However, the training of these models is computationally intensive and susceptible to faults, particularly in the attention mechanism, which is a critical component of transformer-based LLMs. In this paper, we investigate the impact of faults on LLM training, focusing on INF, NaN, and near-INF values in the computation results with systematic fault injection experiments. We observe the propagation patterns of these errors, which can trigger non-trainable states in the model and disrupt training, forcing the procedure to load from checkpoints. To mitigate the impact of these faults, we propose ATTNChecker, the first Algorithm-Based Fault Tolerance (ABFT) technique tailored for the attention mechanism in LLMs. ATTNChecker is designed based on fault propagation patterns of LLM and incorporates performance optimization to adapt to both system reliability and model vulnerability while providing lightweight protection for fast LLM training. Evaluations on four LLMs show that ATTNChecker on average incurs on average 7% overhead on training while detecting and correcting all extreme errors. Compared with the state-of-the-art checkpoint/restore approach, ATTNChecker reduces recovery overhead by up to 49×.

Liang, Yuhang [University of Alabama - Birmingham]↗

Mitigating Fading in Cislunar Communications: Application to the Human Landing System

NASA’s human exploration program is currently working towards landing astronauts on the surface of the Moon by 2024, close the lunar South Pole. To guarantee astronaut safety and maximize science data return, NASA is in the process of defining the communication architecture that will support all astronaut activities from launch to surface operations. Of particular interest to this paper are links from the lunar surface back to Earth without any intermediate relays. We show that the system geometry is such that antennas on the landing system will need to be pointed at low elevation angles, thus potentially causing multi-path fading effects not typically encountered in space communications. This paper is organized in three parts. First, we characterize the multi-path fading effects expected in links between the lunar South Pole and Earth and show that for moderate data rates (less than 1 Mbps) the links suffer from slow fading. We then show that for this operations regime the performance of forward error correction schemes is significantly worse for traditional Additive White Gaussian Noise channels. Finally, we investigate multi-copy mechanisms to mitigate the effects of fading, most notably repetition schemes and Automatic Repeat Request.

Sanchez Net, Marc↗

Enhancing SWAT with mechanistic plant hydraulics: development and application in the Hanjiang River Basin

Plant transpiration plays a critical role in global water and energy cycles, requiring better process understanding as climate change intensifies drought stress and alters plant responses. Most hydrological models such as the widely-used SWAT lack representation of plant hydraulics, the mechanistic processes controlling plant water regulation and transpiration. Here, this study developed SWAT-PHS by integrating a plant hydraulics scheme (PHS) into SWAT hydrological model, enabling explicit simulation of root water uptake, sap flow, storage and transpiration at 30-minute timescales for watershed-scale modeling. In the Hanjiang River Basin, SWAT-PHS mitigated overestimation of runoff during the rainy season and underestimation during the dry season, reducing the overall simulation error by 29% across the entire simulation period. The model can simulate reasonable plant water dynamics, including diurnal transpiration patterns and drought responses showing declining transpiration flux, hydraulic buffering through stem water storage, and depth-dependent root water uptake strategies. Sensitivity analysis shows that SWAT-PHS captured mechanistic relationships between plant hydraulic traits and transpiration, with root distribution and stem capacitance positively affecting annual transpiration while vulnerability parameters showed negative effects. This work provides a pathway for improving hydrologic modeling and water resource management by better representing plant water regulation under climate change and expected intensifying water stress conditions.

China↗

Colloidal nanocrystals for heterogeneous catalysis

Catalytic materials are an essential component of the chemical industry. They find applications in everything from fine chemical manufacturing to greenhouse gas mitigation. They are indispensable for developing a sustainable future. Their development has been continuous, from early trial and error efforts to the first fundamental insights gained through surface science, to modern in-situ characterization and computational predictions. The accumulation of knowledge on the working principles of catalytic surfaces allowed designing and producing better systems with improved performance. Even though tremendous progress has been made thanks to surface science techniques, these studies are usually performed under ultra-high vacuum and are therefore limited in their applicability to more relevant industrial conditions. The control over size, shape and composition in colloidal nanocrystals makes them formidable precursors for model heterogeneous catalysts. These model systems enable linking the insights from surface science studies via in-situ and operando studies to realistic catalytic reaction conditions. In this review, colloidal nanocrystals are presented as powerful building blocks for catalytic materials in the quest for fundamental understanding. A review of the principal methods to produce colloidal nanocrystals with a high level of control is reported, complemented by procedures for how to prepare active catalysts from these particles. Examples and guidelines for the catalytic applications of these materials revolve around the three guiding objectives in catalysis science: activity, selectivity and stability. This work will be limited to examples of this colloidal approach in the areas of thermal, electro- and photocatalysis. In conclusion, the exposed approaches can be used and extended to many other areas of catalysis science, thus providing a new avenue to explore fundamentals and applications of catalytic materials.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗

Upset Characterization of the PowerPC405 Hard-core Processor Embedded in Virtex-II Pro Field Programmable Gate Arrays

Shown in this presentation are recent results for the upset susceptibility of the various types of memory elements in the embedded PowerPC405 in the Xilinx V2P40 FPGA. For critical flight designs where configuration upsets are mitigated effectively through appropriate design triplication and configuration scrubbing, these upsets of processor elements can dominate the system error rate. Data from irradiations with both protons and heavy ions are given and compared using available models.

radiation effects↗

Progress of Coating Stress Compensation of Silicon Mirrors for Lynx X-ray Telescope Mission Concept Using Thermal Oxide Patterning Method

A thermal oxide patterning method has proven to be effective for correcting coating-stress-induced distortion on flat silicon wafers. We report progress on developing this method for correcting curved silicon mirrors distorted by front-side iridium coatings. Owing to the difference in geometry, a finite element model has been established to calculate the appropriate duty cycle maps in thermal oxide hexagon patterns used for compensation. In addition, a photolithographic process, along with three-dimensional printed equipment, has been developed for creating patterns precisely on the back side of curved mirrors. The developed method has been used to recover the original surface shape of two silicon mirrors which are 100-mm long, 0.5-mm thick, having 312-mm radius of curvature, and 30 deg in azimuthal span (Wolter-I geometry). These mirrors’ front sides are sputter-coated by 20-nm iridium layers with ∼-70 N∕m integrated stress. Measurement results show that the developed method can mitigate coating-induced distortion by a factor of ∼5 in RMS height and ∼4 in RMS slope error, corresponding to ∼0.5 arc sec RMS slope error. Residual errors after correction are dominated by midfrequency ripples created during the annealing process, which will be resolved in the future. The presented method is precise and inexpensive and a potential candidate for resolving the coating stress issue for Lynx optics in the future.

x-ray↗

Improved Representation of the Global Dust Cycle Using Observational Constraints on Dust Properties and Abundance

Even though desert dust is the most abundant aerosol by mass in Earth’s atmosphere, atmospheric models 25 struggle to accurately represent its spatial and temporal distribution. These model errors are partially caused byfundamental difficulties in simulating dust emission in coarse-resolution models and in accurately representing dust microphysical properties. Here we mitigate these problems by developing a new methodology that yields an improved representation of the global dust cycle. We present an analytical framework that uses inverse modeling to integrate an ensemble of global model simulations with observational constraints on the dust size distribution, 30 extinction efficiency, and regional dust aerosol optical depth. We then compare the inverse model results against independent measurements of dust surface concentration and deposition flux and find that errors are reduced by approximately a factor of two relative to current model simulations of the Northern Hemisphere dust cycle. The inverse model results show smaller improvements in the less dusty Southern Hemisphere, most likely because both the model simulations and the observational constraints used in the inverse model are less accurate. On a global 35 basis, we find that the emission flux of dust with geometric diameter up to 20 μm (PM20) is approximately 5,000 Tg/year, which is greater than most models account for. This larger PM20 dust flux is needed to match observational constraints showing a large atmospheric loading of coarse dust. We obtain gridded data sets of dust emission, vertically integrated loading, dust aerosol optical depth, (surface) concentration, and wet and dry deposition fluxes that are resolved by season and particle size. As our results indicate that this data set is more accurate than current 40 model simulations and the MERRA-2 dust reanalysis product, it can be used to improve quantifications of dust impacts on the Earth system.

aerosols, dust↗

Planning Bias: Planning as a Source of Sampling Bias

Many data-driven planning methods are trained on data generated by planners. It is well known that many statistical learning methods are sensitive to sampling bias, and yet there has been little or no attention to planning as a sampling method and its role in introducing sampling bias into planner-generated training data. Recently, it has been demonstrated that A**,* in the presence of problems with variable heuristic error, prefers some solutions over other equally cost-optimal solutions. But, as we discuss in this paper, mitigation may not be as simple as resolving arbitrary tie-breaking by sampling from ties uniformly at random. In this paper, we formalize an intuition of planning bias. We focus on problems which output a single solution. Diverse planning only complicates the problem by generalizing it to bias in the set of sets; we show how it is subject to bias in the single solution. We make some useful observations about deterministic algorithms in contrast to non-deterministic algorithms. We explain how information entropy may be a good way to measure planning bias, and discuss some issues in evaluating practical approaches to measurement. We address the intuition that uniform random tiebreaking should mitigate bias; and sketch a novel approach to constructing an appropriate random distribution for duplicate detection during forward search for unbiased A*. Finally, we suggest directions for future work.

Planning Scheduling Algorithms↗

Strong Lensing Cosmology with Population-level Calibrated Neural Ratio Estimation

Strong gravitational lensing contains key information about cosmic acceleration. Modern and next-generation galaxy imaging surveys are expected to provide high-quality data on $\mathcal{O}(10^5)$ galaxy-galaxy lensing systems. The plethora and complexity of the data are likely to present computational challenges for parameter inference methods for fitting high-dimensional likelihoods, which are often analytically intractable. Neural Ratio Estimation (NRE) efficiently computes individual likelihood ratios that can be combined into population-level posteriors. We use simulations to study the capacity of NRE to jointly predict the dark energy equation-of-state parameter $w$ and the total matter density $Ω_{m}$ from lensing images and companion spectroscopic information. We also introduce a post hoc posterior coverage calibration procedure that mitigates the model overconfidence that is typically found in neural density estimation applications. Our experiments show that the errors on both parameters decrease with increasing inference population sizes. In particular, for 100 lenses in a standard $Λ$CDM Universe, our calibrated NRE model achieves median fractional uncertainty of $22.8\%$ in $w$ and $2.9\%$ in $Ω_{m}$. This proof of concept demonstrates a potentially scalable approach for efficient cosmological parameter inference with large populations of galaxy-scale lenses observed in future surveys.

Jarugula, Sreevani [Fermilab] (ORCID:0000000253867↗

QuantumNAT: quantum noise-aware training with noise injection, quantization and normalization

Parameterized Quantum Circuits (PQC) are promising towards quantum advantage on near-term quantum hardware. However, due to the large quantum noises (errors), the performance of PQC models has a severe degradation on real quantum devices. Take Quantum Neural Network (QNN) as an example, the accuracy gap between noise-free simulation and noisy results on IBMQ-Yorktown for MNIST-4 classification is over 60%. Existing noise mitigation methods are general ones without leveraging unique characteristics of PQC; on the other hand, existing PQC work does not consider noise effect. To this end, we present QuantumNAT, a PQC-specific framework to perform noise-aware optimizations in both training and inference stages to improve robustness. We experimentally observe that the effect of quantum noise to PQC measurement outcome is a linear map from noise-free outcome with a scaling and a shift factor. Motivated by that, we propose post-measurement normalization to mitigate the feature distribution differences between noise-free and noisy scenarios. Furthermore, to improve the robustness against noise, we propose noise injection to the training process by inserting quantum error gates to PQC according to realistic noise models of quantum hardware. Lastly, post-measurement quantization is introduced to quantize the measurement outcomes to discrete values, achieving the denoising effect. Extensive experiments on 8 classification tasks using 6 quantum devices demonstrate that QuantumNAT improves accuracy by up to 43%, and achieves over 94% 2-class, 80% 4-class, and 34% 10-class classification accuracy measured on real quantum computers. The code for construction and noise-aware training of PQC is available in the TorchQuantum library.

Wang, Hanrui↗

Sidelobe modeling and mitigation for a three mirror anastigmat cosmic microwave background telescope

Telescopes measuring cosmic microwave background (CMB) polarization on large angular scales require exquisite control of systematic errors to ensure the fidelity of the cosmological results. In particular, far-sidelobe contamination from wide angle scattering is a potentially prominent source of systematic error for large aperture microwave telescopes. Here we describe and demonstrate a ray-tracing-based modeling technique to predict far sidelobes for a three mirror anastigmat telescope designed to observe the CMB from the South Pole. Those sidelobes are produced by light scattered in the receiver optics subsequently interacting with the walls of the surrounding telescope enclosure. After comparing simulated sidelobe maps and angular power spectra for different enclosure wall treatments, we propose a highly scattering surface that would provide more than an order of magnitude reduction in the degree-scale far-sidelobe contrast compared to a typical reflective surface. Finally, we conclude by discussing the fabrication of a prototype scattering wall panel and presenting measurements of its angular scattering profile.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Sidelobe Modeling and Mitigation for A Three Mirror Anastigmat Cosmic Microwave Background Telescope

Telescopes measuring cosmic microwave background (CMB) polarization on large angular scales require exquisite control of systematic errors to ensure the fidelity of the cosmological results. In particular, far-sidelobe contamination from wide angle scattering is a potentially prominent source of systematic error for large aperture microwave telescopes. Here we describe and demonstrate a ray-tracing-based modeling technique to predict far sidelobes for a three mirror anastigmat telescope designed to observe the CMB from the South Pole. Those sidelobes are produced by light scattered in the receiver optics subsequently interacting with the walls of the surrounding telescope enclosure. After comparing simulated sidelobe maps and angular power spectra for different enclosure wall treatments, we propose a highly scattering surface that would provide more than an order of magnitude reduction in the degree-scale far-sidelobe contrast compared to a typical reflective surface. We conclude by discussing the fabrication of a prototype scattering wall panel and presenting measurements of its angular scattering profile.

Ian Gullett↗

Flight Deck Surface Trajectory-based Operations (STBO): Results of Piloted Simulations and Implications for Concepts of Operation (ConOps)

The results offour piloted medium-fidelity simulations investigating flight deck surface trajectory-based operations (STBO) will be reviewed. In these flight deck STBO simulations, commercial transport pilots were given taxi clearances with time and/or speed components and required to taxi to the departing runway or an intermediate traffic intersection. Under a variety of concept of operations (ConOps) and flight deck information conditions, pilots' ability to taxi in compliance with the required time of arrival (RTA) at the designated airport location was measured. ConOps and flight deck information conditions explored included: Availability of taxi clearance speed and elapsed time information; Intermediate RTAs at intermediate time constraint points (e.g., intersection traffic flow points); STBO taxi clearances via ATC voice speed commands or datal ink; and, Availability of flight deck display algorithms to reduce STBO RTA error. Flight Deck Implications. Pilot RTA conformance for STBO clearances, in the form of ATC taxi clearances with associated speed requirements, was found to be relatively poor, unless the pilot is required to follow a precise speed and acceleration/deceleration profile. However, following such a precise speed profile results in inordinate head-down tracking of current ground speed, leading to potentially unsafe operations. Mitigating these results, and providing good taxi RTA performance without the associated safety issues, is a flight deck avionics or electronic flight bag (EFB) solution. Such a solution enables pilots to meet the taxi route RTA without moment-by-moment tracking of ground speed. An avionics or EFB "error-nulling" algorithm allows the pilot to view the STBO information when the pilot determines it is necessary and when workload alloys, thus enabling the pilot to spread his/her attention appropriately and strategically on aircraft separation airport navigation, and the many other flight deck tasks concurrently required. Surface Traffic Management (STM) System Implications. The data indicate a number of implications regarding specific parameters for ATC/STM algorithm development. Pilots have a tendency to arrive at RTA points early with slow required speeds, on time for moderate speeds, and late with faster required speeds. This implies that ATC/STM algorithms should operate with middle-range speeds, similar to that of non-STBO taxi performance. Route length has a related effect: Long taxi routes increase the earliness with slow speeds and the lateness with faster speeds. This is likely due to the" open-loop" nature of the task in which the speed error compounds over a longer time with longer routes. Results showed that this may be mitigated by imposing a small number oftime constraint points each with their own RTAs effectively tuming a long route into a series of shorter routes - and thus improving RTA performance. STBO ConOps Implications. Most important is the impact that these data have for NextGen STM system ConOps development. The results of these experiments imply that it is not reasonable to expect pilots to taxi under a "Full STBO" ConOps in which pilots are expected to be at a predictable (x,y) airport location for every time (t). An STBO ConOps with a small number of intermediate time constraint points and the departing runway, however, is feasible, but only with flight deck equipage enabling the use of a display similar to the "error-nulling algorithm/display" tested.

Foyle, David C.↗

Improved representation of the global dust cycle using observational constraints on dust properties and abundance

Even though desert dust is the most abundant aerosol by mass in Earth's atmosphere, atmospheric models struggle to accurately represent its spatial and temporal distribution. These model errors are partially caused by fundamental difficulties in simulating dust emission in coarse-resolution models and in accurately representing dust microphysical properties. Here we mitigate these problems by developing a new methodology that yields an improved representation of the global dust cycle. We present an analytical framework that uses inverse modeling to integrate an ensemble of global model simulations with observational constraints on the dust size distribution, extinction efficiency, and regional dust aerosol optical depth. We then compare the inverse model results against independent measurements of dust surface concentration and deposition flux and find that errors are reduced by approximately a factor of 2 relative to current model simulations of the Northern Hemisphere dust cycle. The inverse model results show smaller improvements in the less dusty Southern Hemisphere, most likely because both the model simulations and the observational constraints used in the inverse model are less accurate. On a global basis, we find that the emission flux of dust with a geometric diameter up to 20 µm (PM20) is approximately 5000 Tg yr−1, which is greater than most models account for. This larger PM20 dust flux is needed to match observational constraints showing a large atmospheric loading of coarse dust. We obtain gridded datasets of dust emission, vertically integrated loading, dust aerosol optical depth, (surface) concentration, and wet and dry deposition fluxes that are resolved by season and particle size. As our results indicate that this dataset is more accurate than current model simulations and the MERRA-2 dust reanalysis product, it can be used to improve quantifications of dust impacts on the Earth system.

Dust cycle↗

Improving Lidar Turbulence Estimates for Wind Energy

Remote sensing devices (e.g., lidars) are quickly becoming a cost-effective and reliable alternative to meteorological towers for wind energy applications. Although lidars can measure mean wind speeds accurately, these devices measure different values of turbulence intensity (TI) than an instrument on a tower. In response to these issues, a lidar TI error reduction model was recently developed for commercially available lidars. The TI error model first applies physics-based corrections to the lidar measurements, then uses machine-learning techniques to further reduce errors in lidar TI estimates. The model was tested at two sites in the Southern Plains where vertically profiling lidars were collocated with meteorological towers. Results indicate that the model works well under stable conditions but cannot fully mitigate the effects of variance contamination under unstable conditions. To understand how variance contamination affects lidar TI estimates, a new set of equations was derived in previous work to characterize the actual variance measured by a lidar. Terms in these equations were quantified using a lidar simulator and modeled wind field, and the new equations were then implemented into the TI error model.

17 WIND ENERGY↗