Results for “COMPRESSION” — Page 34

DOE OSTI2024

Identifying Outliers in AI-based Image Compression

Image compression using artificial intelligence (AI) is becoming increasingly prevalent across various fields, including scientific research. Scientific instruments can generate hundreds of images per second, and effectively compressing these images with high compression ratios is crucial for facilitating scientific discoveries. However, automatically detecting outlier cases, where compression may not have succeeded or where interesting scientific phenomena are present, poses a significant challenge. To address this, we have developed a methodology based on unsupervised machine learning techniques for detecting outlier compressed images. This methodology utilizes metrics such as peak signal-to-noise ratio (PSNR), structural similarity index measure (SSIM), structural texture similarity index measure (STSIM), and deep image and structural texture similarity index (DISTS). We have evaluated our methodology on several unlabeled datasets, including microscopy and x-ray images, and have successfully identified multiple outlier images using our proposed approach. Furthermore, our approach has enabled us to identify image semantics that are valuable for post-experiment analysis by scientists.

Data Analysis↗

DOE OSTI2023

Online and Scalable Data Compression Pipeline with Guarantees on Quantities of Interest

Data compression is becoming critical for data-intensive scientific applications. Scientists require compression techniques that accurately preserve derived quantities of interest (QoIs). Prior work has shown that a pipeline can be built to guarantee error on the primary data (PD) within user-defined bounds and achieve near-floating point QoI errors. In this paper, we present novel computational approaches for accelerating the pipeline and demonstrate results that enable concurrent execution of compression in parallel with the simulation nodes. This allows compression, including the writing of the required compression data, for the previous time step to be completed while the simulation proceeds with the current time step. Overall, the approach presented in this paper results in a 6–8 times improvement in computational overhead compared to previous work. These results were obtained using data generated by a large-scale fusion code called XGC, which produces hundreds of terabytes of data in a single day.

Banerjee, Tania↗

DOE OSTI2024

Shifting Between Compute and Memory Bounds: A Compression-Enabled Roofline Model

In the evolving landscape of high-performance computing, especially to fight the end of Moore’s Law and Dennard’s Scaling, the ability to shift between compute-bound and memory-bound states is critical for enhancing adaptability and flexibility to diverse system and domain-specific architectures. Such capability is vital for optimizing performance across distinguished hardware configurations, such as accelerators, memory hierarchies, and cache systems. Despite that ad hoc optimization techniques, such as compressed/approximate computation, have been enabled for compute-/data-intensive computing for improved performance in distinct hardware settings, there lacks an understanding of 1) the rational behind performance improvement; 2) capability of different optimizations; 3) what optimization to respond to specific computational and memory demands. This work proposes a compression-enabled roofline model to facilitate this adaptability with data compression techniques to balance and transform between computational and memory demands. This model enables applications to adjust in response to the specific strengths and limitations of the underlying hardware and system to optimize resource utilization. The effectiveness of this approach is demonstrated with matrix multiplication kernels on different input sizes, with turning on/off various compression techniques, including 1) low-precision floating point; 2) sparse matrix formulation; and 3) compressed arrays with ZFP. By reducing memory transfer volumes and cache misses and increasing data locality and computational intensity through compression, the specific roofline model can transform between compute and memory bounds to align more efficiently with system capabilities. This advancement not only improves overall performance but also maximizes adaptability in diverse computing environments.

Naraparaju, Ramasoumya [University of Washington]↗

DOE OSTI2021

Compressive Neural Representations of Volumetric Scalar Fields

Here, we present an approach for compressing volumetric scalar fields using implicit neural representations. Our approach represents a scalar field as a learned function, wherein a neural network maps a point in the domain to an output scalar value. By setting the number of weights of the neural network to be smaller than the input size, we achieve compressed representations of scalar fields, thus framing compression as a type of function approximation. Combined with carefully quantizing network weights, we show that this approach yields highly compact representations that outperform state-of-the-art volume compression approaches. The conceptual simplicity of our approach enables a number of benefits, such as support for time-varying scalar fields, optimizing to preserve spatial gradients, and random-access field evaluation. We study the impact of network design choices on compression performance, highlighting how simple network architectures are effective for a broad range of volumes.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

Numerical Study of Liquid Piston Compression Using Large-Eddy Simulation and Volume-of-Fluid Approach

Efforts to increase the efficiency of residential and commercial air conditioners and heat pumps have demonstrated that the compressor accounts for most of the system’s electrical energy usage. Therefore, the efficiency of this component should be improved to reduce its energy usage. The US Department of Energy’s Oak Ridge National Laboratory developed a near-isothermal liquid piston compressor (LPC) that uses propylene glycol (PG) to compress CO2. This report presents numerical studies of the LPC in which the compression chamber fills with injected PG from the bottom inlet. Numerical simulations were performed using the large-eddy simulation (LES) with the wall-adapting local eddy-viscosity (WALE) subgrid-scale model coupled with the multiphase volume-of-fluid (VOF) model to simulate the transient interface between gas and liquid and to capture the heat and mass transfer within the compression chamber. In this effort, effects of boundary conditions applied to the LES-VOF calculations (e.g., no wall, an adiabatic wall, and a wall with a heat flux subscribed) to the overall pressure and temperature of CO2 gas as well as the transient evolution of flow and heat transfer evolution within the compression chamber are investigated and discussed. It was found that the LES calculation with no wall have shown no dynamical flow patterns and the volume-averaged temperature of CO2 increased from 305 to 392.7 K, while LES calculations with a constant wall temperature or a wall heat flux had similar increases of CO2 temperatures. Results of LES simulation using a wall heat flux showed different stages in the compression process and revealed dynamical formation and interaction of CO2 gas layers and circulation flow patterns within the chamber that contribute to the overall heat transfer between the solid wall, gas, and liquid surface in the compressor.

Nguyen, Thien D.↗

DOE OSTI2016

SU-F-T-121: Abdominal Compression Effectively Reduces the Interplay Effect and Enables Pencil Beam Scanning Proton Therapy of Liver Tumors

Purpose: To study if abdominal compression can reduce breathing motion and mitigate interplay effect in pencil beam scanning proton therapy (PBSPT) treatment of liver tumors in order to better spare healthy liver volumes compared with photon therapy. Methods: Ten patients, six having large tumors initially treated with IMRT and four having small tumors treated with SBRT, were replanned for PBSPT. ITV and beam-specific PTVs based on 4D-CT were used to ensure target coverage in PBSPT. The use of an abdominal compression belt and volumetric repainting was investigated to mitigate the interplay effect between breathing motion and PBSPT dynamic delivery. An in-house Matlab script has been developed to simulate this interplay effect. The dose is computed on each phase individually by sorting all spots according to their simulated delivery timing. The final dose distribution is then obtained by accumulating all dose maps to a reference phase. Results: For equivalent target coverage PBSPT reduced average healthy liver dose by 9.5% of the prescription dose compared with IMRT/SBRT. Abdominal compression of 113.2±42.2 mmHg was effective for all 10 patients and reduced average motion by 2.25 mm. As a result, the average ITV volume decreased from 128.2% to 123.1% of CTV volume. Similarly, the average beam-specific PTV volume decreased from 193.2% to 183.3%. For 8 of the 10 patients, the average motion was reduced below 5 mm, and up to 3 repainting were sufficient to mitigate interplay. For the other two patients with larger residual motion, 4–5 repainting were needed. Conclusion: We recommend evaluation of the 4DCT motion histogram following simulation and the interplay effect following treatment planning in order to personalize the use of compression and volumetric repainting for each patient. Abdominal compression enables safe and more effective PBS treatment of liver tumors by reduction of motion and interplay effect. Kevin Souris is supported by IBA and Televie Grant from F.R.S.-FNRS. Liyong Lin is partially supported by Varian.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2024

ZFP: A compressed array representation for numerical computations

HPC trends favor algorithms and implementations that reduce data motion relative to FLOPS. We investigate the use of lossy compressed data arrays in place of traditional IEEE floating point arrays to store the primary data of calculations. Simulation is fundamentally an exercise in controlled approximation, and error introduced by finite-precision arithmetic (or lossy compression) is just one of several sources of error that need to be managed to ensure sufficient accuracy in a computed result. We describe ZFP, a compressed numerical format designed for in-memory storage of multidimensional arrays, and summarize theoretical results that demonstrate that the error of repeated lossy compression can be bounded and controlled. Furthermore, we establish a relationship between grid resolution and compression-induced errors and show that, contrary to conventional floating point, ZFP reduces finite-difference errors with finer grids. We present example calculations that demonstrate data reduction by 4x or more with negligible impact on solution accuracy. Our results further demonstrate several orders-of-magnitude increase in accuracy using ZFP over IEEE floating point and Posits for the same storage budget.

Lindstrom, Peter↗

DOE OSTI2025

Few-cycle self-compression of GW mid-IR pulses in an anti-resonant fiber in ambient air

We report self-compression of mid-IR optical pulses to the few-cycle regime in an anti-resonant hollow core fiber filled with ambient air. Transform-limited 120-fs duration input optical pulses with 160 μJ energy are compressed to 29.8 fs (< 3 cycles) with a compression efficiency of 75%, resulting in 2.6 GW peak power few-cycle pulses. In addition, we show that nonlinear compression is maintained when the fiber is coiled, allowing for a compact setup. Lastly, we elucidate the dynamics involved in this self-compression setup via numerical simulations and identify the critical role played by Raman shifting in molecular gases and ionization limitations.

Hettel, Will [University of Colorado & NIST, Bould↗

DOE OSTI2022

To what extent can space be compressed? Bandwidth limits of spaceplates

Spaceplates are novel flat-optic devices that implement the optical response of a free-space volume over a smaller length, effectively “compressing space” for light propagation. Together with flat lenses such as metalenses or diffractive lenses, spaceplates have the potential to enable the miniaturization of any free-space optical system. While the fundamental and practical bounds on the performance metrics of flat lenses have been well studied in recent years, a similar understanding of the ultimate limits of spaceplates is lacking, especially regarding the issue of bandwidth, which remains as a crucial roadblock for the adoption of this platform. In this work, we derive fundamental bounds on the bandwidth of spaceplates as a function of their numerical aperture and compression ratio (ratio by which the free-space pathway is compressed). The general form of these bounds is universal and can be applied and specialized for different broad classes of space-compression devices, regardless of their particular implementation. Our findings also offer relevant insights into the physical mechanism at the origin of generic space-compression effects and may guide the design of higher performance spaceplates, opening new opportunities for ultra-compact, monolithic, planar optical systems for a variety of applications.

Shastri, Kunal (ORCID:0000000169282753)↗

DOE OSTI2015

ASTM D395 Short-Term Compression Set of Solid (Non-Porous) Siloxanes: SE 1700, Sylgard 184, and "New" M9787

Compression set of solid (non-porous) Dow Corning SE 1700, Sylgard 184, and “new” M9787 siloxane elastomers was measured according to ASTM D395 Method B. Specimens of SE 1700 were made using (1) the manufacturer’s suggested cure of 150°C for 30 min and (2) an extended cure of 60°C for 6 h and 150°C for 1 h followed by a post-cure under nitrogen purge at 125°C for 12 h. Four specimens of each material were aged at 25-27% compressive strain at 70°C under nitrogen purge for 70 h. Final thickness of each specimen was measured after a 30-min cooling/relaxation period, and compression set relative to deflection was calculated. The average compression set relative to deflection was 6.0% for SE 1700 made using the extended cure and post-cure, 11.3% for SE 1700 made using the manufacturer’s suggested cure, 12.1% for Sylgard 184, and 1.9% for M9787. The extended cure and post-cure reduced the amount of compression set in SE 1700.

36 MATERIALS SCIENCE↗

DOE OSTI2015

Geothermally Coupled Well-Based Compressed Air Energy Storage

Previous work by McGrail et al. (2013, 2015) has evaluated the possibility of pairing compressed air energy storage with geothermal resources in lieu of a fossil-fired power generation component, and suggests that such applications may be cost competitive where geology is favorable to siting both the geothermal and CAES components of such a system. Those studies also note that the collocation of subsurface resources that meet both sets of requirements are difficult to find in areas that also offer infrastructure and near- to mid-term market demand for energy storage. This study examines a novel application for the compressed air storage portion of the project by evaluating the potential to store compressed air in disused wells by amending well casings to serve as subsurface pressure vessels. Because the wells themselves would function in lieu of a geologic storage reservoir for the CAES element of the project, siting could focus on locations with suitable geothermal resources, as long as there was also existing wellfield infrastructure that could be repurposed for air storage. Existing wellfields abound in the United States, and with current low energy prices, many recently productive fields are now shut in. Should energy prices remain stagnant, these idle fields will be prime candidates for decommissioning unless they can be transitioned to other uses, such as redevelopment for energy storage. In addition to the nation’s ubiquitous oil and gas fields, geothermal fields, because of their phased production lifetimes, also may offer many abandoned wellbores that could be used for other purposes, often near currently productive geothermal resources. These existing fields offer an opportunity to decrease exploration and development uncertainty by leveraging data developed during prior field characterization, drilling, and production. They may also offer lower-cost deployment options for hybrid geothermal systems via redevelopment of existing well-field infrastructure. This project assessed the technical and economic feasibility of implementing geothermally coupled well-based CAES for grid-scale energy storage. Based on an evaluation of design specifications for a range of casing grades common in U.S. oil and gas fields, a 5-MW CAES project could be supported by twenty to twenty-five 5,000-foot, 7-inch wells using lower-grade casing, and as few as eight such wells for higher-end casing grades. Using this information, along with data on geothermal resources, well density, and potential future markets for energy storage systems, The Geysers geothermal field was selected to parameterize a case study to evaluate the potential match between the proven geothermal resource present at The Geysers and the field’s existing well infrastructure. Based on calculated wellbore compressed air mass, the study shows that a single average geothermal production well could provide enough geothermal energy to support a 15.4-MW (gross) power generation facility using 34 to 35 geothermal wells repurposed for compressed air storage, resulting in a simplified levelized cost of electricity (sLCOE) estimated at 11.2 ¢/kWh (Table S.1). Accounting for the power loss to the geothermal power project associated with diverting geothermal resources for air heating results in a net 2-MW decrease in generation capacity, increasing the CAES project’s sLCOE by 1.8 ¢/kWh.

15 GEOTHERMAL ENERGY↗

DOE OSTI2015

Geothermally Coupled Well-Based Compressed Air Energy Storage

Previous work by McGrail et al. (2013, 2015) has evaluated the possibility of pairing compressed air energy storage with geothermal resources in lieu of a fossil-fired power generation component, and suggests that such applications may be cost competitive where geology is favorable to siting both the geothermal and CAES components of such a system. Those studies also note that the collocation of subsurface resources that meet both sets of requirements are difficult to find in areas that also offer infrastructure and near- to mid-term market demand for energy storage. This study examines a novel application for the compressed air storage portion of the project by evaluating the potential to store compressed air in disused wells by amending well casings to serve as subsurface pressure vessels. Because the wells themselves would function in lieu of a geologic storage reservoir for the CAES element of the project, siting could focus on locations with suitable geothermal resources, as long as there was also existing wellfield infrastructure that could be repurposed for air storage. Existing wellfields abound in the United States, and with current low energy prices, many recently productive fields are now shut in. Should energy prices remain stagnant, these idle fields will be prime candidates for decommissioning unless they can be transitioned to other uses, such as redevelopment for energy storage. In addition to the nation’s ubiquitous oil and gas fields, geothermal fields, because of their phased production lifetimes, also may offer many abandoned wellbores that could be used for other purposes, often near currently productive geothermal resources. These existing fields offer an opportunity to decrease exploration and development uncertainty by leveraging data developed during prior field characterization, drilling, and production. They may also offer lower-cost deployment options for hybrid geothermal systems via redevelopment of existing well-field infrastructure. This project assessed the technical and economic feasibility of implementing geothermally coupled well-based CAES for grid-scale energy storage. Based on an evaluation of design specifications for a range of casing grades common in U.S. oil and gas fields, a 5-MW CAES project could be supported by twenty to twenty-five 5,000-foot, 7-inch wells using lower-grade casing, and as few as eight such wells for higher-end casing grades. Using this information, along with data on geothermal resources, well density, and potential future markets for energy storage systems, The Geysers geothermal field was selected to parameterize a case study to evaluate the potential match between the proven geothermal resource present at The Geysers and the field’s existing well infrastructure. Based on calculated wellbore compressed air mass, the study shows that a single average geothermal production well could provide enough geothermal energy to support a 15.4-MW (gross) power generation facility using 34 to 35 geothermal wells repurposed for compressed air storage, resulting in a simplified levelized cost of electricity (sLCOE) estimated at 11.2 ¢/kWh (Table S.1). Accounting for the power loss to the geothermal power project associated with diverting geothermal resources for air heating results in a net 2-MW decrease in generation capacity, increasing the CAES project’s sLCOE by 1.8 ¢/kWh.

25 ENERGY STORAGE↗

DOE OSTI2015

Pre-Compressed Targets for Gas Gun Studies

This work developed the capability for pre-compressing targets for Hugoniot measurements utilizing gas gun driven flyer plates. Initial condition control, such as pre-compression, allows access to states off the principal Hugoniot and in the case of highly compressible materials, the locus of shock states may be well off (significantly denser than) the principal Hugoniot even for relatively modest (~100's MPa) initial pressures. The design and operation of the pre-compression target hardware will be discussed in relation to functionality and design considerations for dynamic testing. Capability for loading of super critical gases and gas mixtures has been successfully demonstrated and dynamic hypervelocity impacts of helium were tested. Example loading of supercritical alcohol (ethanol) will is presented in the context of previous impact studies of initially ambient quiescent liquid. Advantages and limitations of the system are discussed in the context accessing highly compressed states.

42 ENGINEERING↗

DOE OSTI2019

Earth Battery: Storing Energy with Compressed Air and Heated Brine in Porous Rock (Final Technical Report)

Cost-effective reductions in greenhouse gas (GHG) emissions are best achieved when all lowcarbon energy resources are fully utilized. This includes base-load power: nuclear energy (NE) and fossil energy (FE), integrated with CO 2 capture, use, and storage (CCUS), and variable renewable energy (VRE) (wind and solar); but, current CCUS options don’t justify CO 2 -capture costs and existing energy-storage approaches lack the capacity and storage duration needed to fully utilize all forms of low-carbon energy without curtailment. The Earth Battery is designed to meet these challenges by synergistically integrating conventional- and renewable-energy resources, allowing each to contribute more efficiently to the grid than if operated independently. The Earth Battery stores energy underground as pressure and heat, using compressed air and/or CO 2 , together with heated brine. These fluids are stored in either saline aquifers used for CO 2 or natural gas (NG) storage or in oil and gas reservoirs, making deployment possible over much of the U.S. We can provide enormous energy-storage capacity and duration. The main goal of this study was to assess the techno-economic feasibility of the versions of the Earth Battery that use compressed air energy storage (CAES). Current CAES systems use NG turbines and store air in salt caverns, which limits geographic deployment. Conventional CAES wastes much of the heat of air compression, comprising half of the compression energy. Our technology addresses these deficiencies by using permeable sedimentary rock to store air and the heat of compression as heated brine. When electricity is needed, air and hot brine are produced, with hot brine used to pre-heat air before it is fed into the expanders. Key factors affecting round-trip efficiency are (1) pressure loss in the air-storage reservoir, which depends on permeability (> 100 mD being needed for efficient operations) and (2) temperature loss in the hot-brine storage reservoir, which decreases with time as the storage formation heats up.

25 ENERGY STORAGE↗

DOE OSTI2021

(U) Validation and Optimization of a 1-Dimensional Quasi-Isentropic Compression Model Using a Functionally Graded Material Flyer Plate

The fields studying highly compressed materials have been around for many years. These fields have given birth to better understandings of the solar fusion cycle and the liquid-metallic core of Jupiter. The highest pressures achieved have been at the National Ignition Facility (NIF) and more recently, the Z-Machine. One method to achieve higher levels of compression is to use quasi-isentropic (QI) compression. In practice, this means minimizing the amount of temperature increase of the target material. In their paper, J.H. Nguyen et. al. manufactured a functionally graded material (FGM) in order to help them achieve QI compression. The aim of the study described in this paper is to validate a 1-dimensional (1D) FLAG model that uses an FGM flyer to compress a copper (Cu) target. Once validated, the density/composition profile of the FGM is varied and optimized using Markov Chain Monte Carlo, specifically using a Metropolis-Hastings algorithm to find the optimal FGM profile that minimizes the temperature increase in the copper.

36 MATERIALS SCIENCE↗

DOE OSTI2022

Compressed Natural Gas Component Leak Frequency Estimation

The frequency of unintended releases in a compressed natural gas system is an important aspect of the system quantitative risk assessment. The frequencies for possible release scenarios, along with engineering models, are utilized to quantify the risks for compressed natural gas facilities. This report documents component leakage frequencies representative of compressed natural gas components that were estimated as a function of the normalized leak size. A Bayesian statistical method was used which results in leak frequency distributions for each component which represent variation and uncertainty in the leak frequency. The analysis shows that there is high uncertainty in the estimated leak frequencies due to sparsity in compressed natural gas data. These leak frequencies may still be useful in compressed natural gas system risk assessments, as long as this high uncertainty is acknowledged and considered appropriately.

03 NATURAL GAS↗

DOE OSTI2024

Production of Readily Compressible Dies for the Enhanced Sintering of Solids (PRESS) Preliminary Investigations

Non-oxide ceramics, which consist of predominantly carbides, borides, and nitrides, are of great interest to modern engineering because of their resistance to extreme conditions. Properties of such materials include high temperature resistance, high hardness, chemical resistance, and high fracture toughness. However, the main method for manufacturing parts of these materials is by machining cylindrical billets to the desired geometry. Not only does this produce waste, but the cost of machining is high since the part materials are of such high hardness. This method is used because these non-oxide ceramics must be sintered to high density using simultaneous heat and pressure within a hot-pressing unit. These units are inherently restricted to cylindrical geometries. A proposed way to expand the capabilities of a hot-press is by using intermediate compressible dies. This method consists of surrounding a ceramic part green body with a material that will shrink at the same rate as the part and survive the hot-pressing conditions. At the end, the compressible die material would be removed leaving a nearly net shaped hot-pressed part. In this study, graphite powder was investigated as the compressible die material with its shrinkage rate being controlled by particle size mixing. Both ceramic materials and graphite were then cast into parts and compressible dies respectively to show the feasibility of the compressible die hot-pressing method.

36 MATERIALS SCIENCE↗

DOE OSTI2025

Automatic Generation of Algorithms for High-Speed Reliable Lossy Data Compression (Final Report)

Fast reliable data compression is urgently needed for many leading-edge scientific instruments and for exascale high-performance computing applications because they produce vast amounts of data at extremely high rates. The goal of this project has been to develop a framework named LC that is able to automatically generate high-speed lossless and reliable lossy compression and decompression algorithms that can be customized for different kinds of data. The resulting LC framework is freely available on GitHub. To achieve high-speed operation, LC outputs optimized and parallelized CPU and GPU implementations of the generated algorithms. To ensure the quality of lossily compressed data, LC guarantees the user-provided error bound. To be able to customize the compression algorithm to various use cases, LC can synthesize millions of different algorithms and automatically search for the one that works best for the given data. We have already employed LC to create state-of-the-art lossless and lossy compressors for scientific data as well as leading lossless compressors for images. We hope that LC and the customized, fast, reliable, and CPU/GPU-compatible compression algorithms that it can generate will greatly benefit the many scientific applications that need not only high trustworthiness but also high performance.

97 MATHEMATICS AND COMPUTING↗