Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “APPLICATIONS”

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 1,495 records · Page 83

miniGAN: a proxy application for generative adversarial networks

miniGAN is a python-based machine learning proxy application for generative adversarial networks, developed through the Exascale Computing Project's (ECP) ExaLearn project. It will be included in the main ECP proxy application and the machine learning proxy application suite. It is a proxy for ECP cosmological(CosmoFlow, ExaGAN) and wind energy(ExaWind) applications. miniGAN will be distributed to ECP hardware vendors as part of hardware codesign. miniGAN uses the Numpy/PyTorch/TensorFlow/Keras/Horovod frameworks and libraries. It also relies on the Kokkos and Kokkos-Kernels packages developed here at Sandia Labs. SAND2020-2038 M Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy's National Nuclear Security Administration under contract DE-NA0003525.

Ellis, John↗

DLIO: A DATA-CENTRIC BENCHMARK FOR DEEP LEARNING APPLICATIONS

SF-22-136 Deep learning has been shown as a successful method for various tasks, and its popularity results in numerous open-source deep learning software tools. Deep learning has been applied to a broad spectrum of scientific domains such as cosmology, particle physics, computer vision, fusion, and astrophysics. Scientists have performed a great deal of work to optimize the computational performance of deep learning frameworks. However, the same cannot be said for I/O performance. As deep learning algorithms rely on big-data volume and variety to effectively train neural networks accurately, I/O is a significant bottleneck on large-scale distributed deep learning training. DLIO, is a novel representative benchmark suite built based on the I/O profiling of the selected workloads. DLIO can be utilized to accurately emulate the I/O behavior of modern deep learning applications. Using DLIO, application developers and system software solution architects can identify potential I/O bottlenecks in their applications and guide optimizations to boost the I/O performance leading to lower training times. The storage vendor can also use DLIO as a guide for designing and optimize the storage and filesystem targeting at deep learning application.

ZHENG, HUIHUO↗

MiniApps derived from production HPC applications using multiple programing models

Here, we have developed a set of reduced, proxy applications (“MiniApps”) based on large-scale application codes supported at the Oak Ridge Leadership Computing Facility (OLCF). The MiniApps are designed to encapsulate the details of the most important (i.e. the most time-consuming and/or unique) facets of the applications that run in production mode on the OLCF. In each case, we have produced or plan to produce individual versions of the MiniApps using different specific programing models (e.g., OpenACC, CUDA, OpenMP). We describe some of our initial observations regarding these different implementations along with estimates of how closely the MiniApps track the actual performance characteristics (in particular, the overall scalability) of the large-scale applications from which they are derived.

97 MATHEMATICS AND COMPUTING↗

AMReX: Block-structured adaptive mesh refinement for multiphysics applications

Block-structured adaptive mesh refinement (AMR) provides the basis for the temporal and spatial discretization strategy for a number of Exascale Computing Project applications in the areas of accelerator design, additive manufacturing, astrophysics, combustion, cosmology, multiphase flow, and wind plant modeling. AMReX is a software framework that provides a unified infrastructure with the functionality needed for these and other AMR applications to be able to effectively and efficiently utilize machines from laptops to exascale architectures. AMR reduces the computational cost and memory footprint compared to a uniform mesh while preserving accurate descriptions of different physical processes in complex multiphysics algorithms. AMReX supports algorithms that solve systems of partial differential equations in simple or complex geometries and those that use particles and/or particle–mesh operations to represent component physical processes. In this article, we will discuss the core elements of the AMReX framework such as data containers and iterators as well as several specialized operations to meet the needs of the application projects. In addition, we will highlight the strategy that the AMReX team is pursuing to achieve highly performant code across a range of accelerator-based architectures for a variety of different applications.

Zhang, Weiqun↗

EXAGRAPH: Graph and combinatorial methods for enabling exascale applications

Combinatorial algorithms in general and graph algorithms in particular play a critical enabling role in numerous scientific applications. However, the irregular memory access nature of these algorithms makes them one of the hardest algorithmic kernels to implement on parallel systems. With tens of billions of hardware threads and deep memory hierarchies, the exascale computing systems in particular pose extreme challenges in scaling graph algorithms. The codesign center on combinatorial algorithms, ExaGraph, was established to design and develop methods and techniques for efficient implementation of key combinatorial (graph) algorithms chosen from a diverse set of exascale applications. Algebraic and combinatorial methods have a complementary role in the advancement of computational science and engineering, including playing an enabling role on each other. In this paper, we survey the algorithmic and software development activities performed under the auspices of ExaGraph from both a combinatorial and an algebraic perspective. In particular, we detail our recent efforts in porting the algorithms to manycore accelerator (GPU) architectures. We also provide a brief survey of the applications that have benefited from the scalable implementations of different combinatorial algorithms to enable scientific discovery at scale. We believe that several applications will benefit from the algorithmic and software tools developed by the ExaGraph team.

97 MATHEMATICS AND COMPUTING↗

The ECP SICM project: Managing complex memory hierarchies for exascale applications

The Exascale Computing Project (ECP)’s Simplified Interface to Complex Memories (SICM) effort focuses on developing universal interfaces for discovering, managing, and sharing data across complex memory hierarchies. These facilitate the exploitation of emerging memory technologies and support precise control over their various trade-offs such as high-bandwidth versus low-latency, persistent versus ephemeral, high-capacity versus low-capacity, and near-CPU versus near-GPU. SICM comprises three interrelated components: a low-level interface, a high-level interface, and a persistent-heap interface. The low-level SICM interface is intended for system and run-time developers as well as expert application developers who prefer full control of the memory objects used within their application. The high-level SICM interface builds upon the low-level interface, employing application-level profiling and analysis to optimize data management for complex memory hierarchies. The persistent-heap interface provides applications with a persistent memory allocator that can allocate custom C++ data structures in both block-storage and byte-addressable persistent memories.

97 MATHEMATICS AND COMPUTING↗

HDF5 in the exascale era: Delivering efficient and scalable parallel I/O for exascale applications

Accurately modeling real-world systems requires scientific applications at exascale to generate massive amounts of data and manage data storage efficiently. However, parallel input and output (I/O) faces challenges due to new application workflows and the state-of-the-art memory, interconnect, and storage architectures considered in exascale designs. The storage hierarchy has expanded with node-local persistent memory, solid-state storage, and traditional disk and tape-based storage, thus requiring efficiency at each layer and much more efficient data movement among these layers. This paper discusses how the ExaHDF5 project improved the I/O performance and data management for exascale architectures by enhancing HDF5, a widely used parallel I/O library. The team developed an Asynchronous I/O Virtual Object Layer (VOL) connector that allowed overlapping I/O with computation. They also created a Cache VOL to complement asynchronous I/O by incorporating fast storage layers, such as burst buffer and node-local storage, into the parallel I/O workflow through caching and staging data. Additionally, the team enabled data aggregation and I/O at the node level by using a Subfiling Virtual File Driver (VFD). To demonstrate superior I/O performance with HDF5 at exascale, the ExaHDF5 team collaborated with several exascale applications. In this paper, we show I/O performance improvements for three applications: Cabana (a particle-based simulation library), EQSIM (a regional earthquake simulation software), and E3SM (a climate system modeling library).

Asynchronous I/Ol↗

3D-printed mirror-less helicity preserving metasurface “mirror” for THz applications

Stimulated by seminal works on generalized reflection and refraction laws, metasurfaces have evolved to a highly promising research direction, as they allow a multitude of different functionalities by optically thin wave-control elements/structures. Among them, structures functioning in the low THz regime are of extensive research interest, due to their high potential in communication and sensing applications. In this paper we propose a simple THz metasurface design that exhibits ideally perfect cross-polarized reflection for linear polarization and perfect helicity-preserving reflection with a geometric (Pancharatnam–Berry) phase for circular polarization, without the presence of any back-reflector. Numerical calculations demonstrating the structure response are justified by analytical models, which provide physical insights on this response. The designed metasurfaces are fabricated using the direct laser writing 3D-printing technology, metallized with electroless silver plating, and are characterized by THz time domain spectroscopy, with the experimental results validating the corresponding theoretical ones. Applications including beam steering and focusing, exploring the Pancharatnam–Berry phase, are also demonstrated numerically. Besides those applications, the helicity preserving mirror response of our metasurfaces can be valuable, among others, also in molecular chirality sensing applications, an issue that is also highlighted here.

3D-printed metasurfaces↗

Nanostructured cerium oxide: preparation, characterization, and application in energy and environmental catalysis

Nanostructured cerium oxide (CeO 2 ) with outstanding physical and chemical properties has attracted extensive interests over the past few decades in environment and energy-related applications. With controllable synthesis of nanostructured CeO 2 , much more features were technologically brought out from defect chemistry to structure-derived effects. This paper highlights recent progress on the synthesis and characterization of nanostructured ceria-based materials as well as the traditional and new applications. Specifically, several typical applications based on the desired ceria nanostructures are focused to showcase the importance of nanostructure-derived effects. Moreover, some challenges and perspectives on the nanostructured ceria are presented, such as defects controlling and retainment, scale-up fabrication, and monolithic devices. Hopefully, this paper can provide an improved understanding of nanostructured CeO 2 and offer new opportunities to promote the further research and applications in the future.

36 MATERIALS SCIENCE↗

Recent advances in metal hydrides for clean energy applications

Metal hydrides are a fascinating class of materials that can be utilized for a surprising variety of clean energy applications, including smart solar collectors, smart windows, sensors, thermal energy storage, and batteries, in addition to their traditional application for hydrogen storage. Over the past decade, research on metal hydrides for hydrogen storage increased due to global governmental incentives and an increased focus on hydrogen storage research for polymer electrolyte membrane fuel cell operation. Tremendous progress has been made in so-called complex metal hydrides for hydrogen storage applications with the discovery of many new hydrides containing covalently bound complex anions. Many of these materials have applications beyond hydrogen storage and are being investigated for lithium-ion battery separator and anode materials. In this issue of MRS Bulletin , we present the state of the art of key evolving metal-hydride-based clean energy technologies with an outlook toward future needs.

batteries↗

Redesign of the Timeline Generator at Fermilab using a web-based Flutter application, GraphQL API and an IOC

The control system at Fermilab is undergoing an evolution with a shift towards web-based applications with connections to the EPICS infrastructure. The Timeline Generator (TLG) is an application that serves to coordinate events across the lab using different timing links. These links include the Tevatron clock (TCLK), a 10 MHz serial link with events encoded at 20Hz and Machine Data (MDAT), a communication link with states encoded at 720Hz. This paper covers the redesign of the major components of the TLG. This includes a web-based Flutter application for building timelines. A placement service is in use that has a GraphQL interface and uses a timeline input to compute a schedule of events and states. The Flutter application sends this computed schedule to the TLG IOC via a GraphQL interface to the Data Pool Manager (DPM). The TLG IOC runs on an Arria FPGA, the Accelerator Clock Generator (ACLK-GEN), which is responsible for writing the events and states on to the different timing links.

Carmichael, Linden [Fermilab]↗

Compact, High-Power Superconducting Electron Linear Accelerators for MW Industrial Applications

Fermilab has developed a novel concept for an industrial electron linac using Nb₃Sn coating technology and conduction cooling. We will show the range of multi-cavity linac designs targeted toward various applications. We will also discuss technology development status with results on conduction cooling of SRF cavities based on cryocoolers, which removes the need for liquid Helium, thus making SRF technology accessible to industrial applications. These conduction-cooled linacs can generate electron beam energies up to 10 MeV in continuous-wave operation and can reach higher power (>=1 MW) by combing several modules. Compact and light enough to mount on mobile platforms, our machine is anticipated to enable new in-situ environmental remediation applications such as waste-water treatment for urban areas, X-ray medical device sterilization, and innovative pavement applications. We also show cost-economics and key R&D areas that much be addressed for a practical machine.

43 PARTICLE ACCELERATORS↗

LDRD final report : managing shared memory data distribution in hybrid HPC applications.

MPI is the dominant programming model for distributed memory parallel computers, and is often used as the intra-node programming model on multi-core compute nodes. However, application developers are increasingly turning to hybrid models that use threading within a node and MPI between nodes. In contrast to MPI, most current threaded models do not require application developers to deal explicitly with data locality. With increasing core counts and deeper NUMA hierarchies seen in the upcoming LANL/SNL 'Cielo' capability supercomputer, data distribution poses an upper boundary on intra-node scalability within threaded applications. Data locality therefore has to be identified at runtime using static memory allocation policies such as first-touch or next-touch, or specified by the application user at launch time. We evaluate several existing techniques for managing data distribution using micro-benchmarks on an AMD 'Magny-Cours' system with 24 cores among 4 NUMA domains and argue for the adoption of a dynamic runtime system implemented at the kernel level, employing a novel page table replication scheme to gather per-NUMA domain memory access traces.

99 GENERAL AND MISCELLANEOUS↗

Optika : a GUI framework for parameterized applications.

In the field of scientific computing there are many specialized programs designed for specific applications in areas such as biology, chemistry, and physics. These applications are often very powerful and extraordinarily useful in their respective domains. However, some suffer from a common problem: a non-intuitive, poorly-designed user interface. The purpose of Optika is to address this problem and provide a simple, viable solution. Using only a list of parameters passed to it, Optika can dynamically generate a GUI. This allows the user to specify parameters values in a fashion that is much more intuitive than the traditional 'input decks' used by some parameterized scientific applications. By leveraging the power of Optika, these scientific applications will become more accessible and thus allow their designers to reach a much wider audience while requiring minimal extra development effort.

99 GENERAL AND MISCELLANEOUS↗

Nuclear Resonance Fluorescence for Safeguards Applications

In nuclear resonance fluorescence (NRF) measurements, resonances are excited by an external photon beam leading to the emission of {gamma} rays with specific energies that are characteristic of the emitting isotope. The promise of NRF as a non-destructive analysis technique (NDA) in safeguards applications lies in its potential to directly quantify a specific isotope in an assay target without the need for unfolding the combined responses of several fissile isotopes as often required by other NDA methods. The use of NRF for detection of sensitive nuclear materials and other contraband has been researched in the past. In the safeguards applications considered here one has to go beyond mere detection and precisely quantify the isotopic content, a challenge that is discussed throughout this report. Basic NRF measurement methods, instrumentation, and the analytical calculation of NRF signal strengths are described in Section 2. Well understood modeling and simulation tools are needed for assessing the potential of NRF for safeguards and for designing measurement systems. All our simulations were performed with the radiation transport code MCNPX, a code that is widely used in the safeguards community. Our initial studies showed that MCNPX grossly underestimated the elastically scattered background at backwards angles due to an incorrect treatment of Rayleigh scattering. While new, corrected calculations based on ENDF form factors showed much better agreement with experimental data for the elastic scattering of photons on an uranium target, the elastic backscatter is still not rigorously treated. Photonuclear scattering processes (nuclear Thomson, Delbruck and Giant Dipole Resonance scattering), which are expected to play an important role at higher energies, are not yet included. These missing elastic scattering contributions were studied and their importance evaluated evaluated against data found in the literature as discussed in Section 3. A transmission experiment was performed in September 2009 to test and demonstrate the applicability of the method to the quantitative measurement of an isotope of interest embedded in a thick target. The experiment, data analysis, and results are described in Section 4. The broad goal of our NRF studies is to assess the potential of the technique in safeguards applications. Three examples are analyzed in Section 5: the isotopic assay of spent nuclear fuel (SNF), the measurement of {sup 235}U enrichment in UF{sub 6} cylinders, and the determination of {sup 239}Pu in mixed oxide (MOX) fuel. The study of NRF for the assay of SNF assemblies was supported by the Next Generation Safeguards Initiative (NGSI) of the U.S. Department of Energy as part of a large multi-lab/university effort to quantify the plutonium (Pu) mass in spent nuclear fuel assemblies and to detect the diversion of pins with non-destructive assay (NDA) methods. NRF is one of 14 NDA techniques being researched. The methodology for performing and analyzing quantitative NRF measurements was developed for determining Pu mass in SNF and is extensively discussed in this report. The same methodology was applied to the assessment of NRF for the measurement of {sup 235}U enrichment and the determination of {sup 239}Pu in MOX fuel. The analysis centers on determining suitable NRF measurement methods, measurement capabilities that could be realized with currently available instrumentation, and photon source and detector requirements for achieving useful NDA capabilities.

98 NUCLEAR DISARMAMENT, SAFEGUARDS, AND PHYSICAL P↗

An Overview of Nuclear vs. Non-Nuclear Design Code Requirements for a Candidate Steam Supply System for Commercial Applications

The objective is to identify (mostly for industrial end-users) the difference between a Section III nuclear steam generator (classified as Structures, Systems and Components (SSC)) and a Section VIII steam generator in the same general conditions, but used in a conventional application. Specifically, applicable temperature and pressure ranges and a more quantitative description of how materials change, design margins change and required design rigor changes are of interest. This overview focuses on the steam generator pressure boundary but the downstream piping will also be considered. Within the designations of Section III and Section VIII there are subcategories with their specific regions of applicability. Each of these subcategories has evolved their own unique features with respect to design rules and their implementation. A general overview of the various design codes will be provided in sufficient detail to illustrate the major differences; however, a detailed discussion of the various design requirements and their implementation is beyond the scope of this discussion. References (1) and (2) are sources of more detailed information. Also, example wall sizing calculations will be provided to illustrate the application of the relevant design codes under the candidate design conditions. The candidate steam supply Design Conditions are 600C (1112F) and 24MPa (3,480psi). The Operating Conditions or Service Levels will be somewhat lower and the difference shows up in some of the various design methodologies employed.

42 ENGINEERING↗

"Dedicated To The Continued Education, Training and Demonstration of PEM Fuel Cell Powered Lift Trucks In Real-World Applications."

The project objective was to further assist in the commercialization of fuel cell and H2 technology by building further upon the successful fuel cell lift truck deployments that were executed by LiftOne in 2007, with longer deployments of this technology in real-world applications. We involved facilities management, operators, maintenance personnel, safety groups, and Authorities Having Jurisdiction. LiftOne strived to educate a broad group from many areas of industry and the community as to the benefits of this technology. Included were First Responders from the local areas. We conducted month long deployments with end-users to validate the value proposition and the market requirements for fuel cell powered lift trucks. Management, lift truck operators, Authorities Having Jurisdiction and the general public experienced 'hands on' fuel cell experience in the material handling applications. We partnered with Hydrogenics in the execution of the deployment segment of the program. Air Products supplied the compressed H2 gas and the mobile fueler. Data from the Fuel Cell Power Packs and the mobile fueler was sent to the DOE and NREL as required. Also, LiftOne conducted the H2 Education Seminars on a rotating basis at their locations for lift trucks users and for other selected segments of the community over the project's 36 month duration. Executive Summary The technology employed during the deployments program was not new, as the equipment had been used in several previous demos and early adoptions within the material handling industry. This was the case with the new HyPx Series PEM - Fuel Cell Power Packs used, which had been demo'd before during the 2007 Greater Columbia Fuel Cell Challenge. The Air Products HF-150 Fueler was used outdoors during the deployments and had similarly been used for many previous demo programs. The methods used centered on providing this technology as the power for electric sit-down lift trucks at high profile companies operating large fleets. As a long-standing lift truck dealership, LiftOne was able to introduce the fuel cells to such companies in the demanding applications. Accomplishments vs Objectives: We were successful in respect to the stated objectives. The Education Segment's H2 Education Sessions were able to introduce fuel cell technology to many companies and reached the intended broad audience. Also, demos of the lift truck at the sessions as well as the conferences; expos and area events provided great additional exposure. The Deployments were successful in allowing the 6 participating companies to test the 2 fuel cell powered lift trucks in their demanding applications. One of the 6 sites (BMW) eventually adopted over 80 fuel cells from Plug Power. LiftOne was one of the 3 fuel cell demonstrators at BMW for this trial and played a major role in helping to prove the viability and efficiency of this alternative form of energy for BMW. The other 5 companies that participated in the project's deployments were encouraged by the trials and while not converting over to fuel cell power at this time, expressed the desire to revisit acquisition scenarios in the near future as the cost of fuel cells and infrastructure continue to improve. The Education sessions began in March of 2009 at the 7 LiftOne Branches and continued throughout the duration of the project. Attendees came from a large base of lift truck users in North Carolina, South Carolina and Virginia. The sessions were free and invitations were sent out to potential users and companies with intrigue. In addition to the Education content at the sessions (which was offered in a 'H2 101' format), LiftOne was able to demonstrate a working fuel cell powered lift truck, which proved to be a big draw with the 'hands on' experience. LiftOne also demo'd the fuel cell lift trucks at many conferences, expos, professional association meetings, trade shows and 'Green' events in major cities region including Charlotte, Greenville, and Columbia. Such events allowed for H2 Education Material to be presented, and recruit attendees for future sessions. The Deployments began in May '09 and continued through August'10. While the overall results were good, there were some technical problems with the fuel cells at times during the trials. These items are detailed in the attachment provided with this submission. There were never any safety issues during the trials as design features and diagnostic capabilities allowed for quick analysis and field repairs. There were no products developed during this award, as we demo'd existing products. During the project, Tom Dever was able to speak at several panel type discussions and presentations at the NHA/ Fuel Cell & H2 Energy Annual Conference, as well as at the Promat Show in Chicago. Expertise and deployment/ education experiences were imparted during these events. The presentations are included as attachments.

08 HYDROGEN↗

Machining Elimination through Application of Thread Forming Fasteners in Net-Shaped Cast Holes

The ultimate objective of this work was to eliminate approximately 30% of the machining performed in typical automotive engine and transmission plants by using thread forming fasteners in as-cast holes of aluminum and magnesium cast components. The primary issues at the source of engineers reluctance to implementing thread forming fasteners in lightweight castings are: • Little proof of consistency of clamp load vs. input torque in either aluminum or magnesium castings. • No known data to understand the effect on consistency of clamp load as casting dies wear. The clamp load consistency concern is founded in the fact that a portion of the input torque used to create clamp load is also used to create threads. The torque used for thread forming may not be consistent due to variations in casting material, hole size and shape due to tooling wear and process variation (thermal and mechanical). There is little data available to understand the magnitude of this concern or to form the basis of potential solutions if the range of clamp load variation is very high (> +/- 30%). The range of variation that can be expected in as-cast hole size and shape over the full life cycle of a high pressure die casting die was established in previous work completed by Pacific Northwest National Laboratory, (PNNL). This established range of variation was captured in a set of 12 cast bosses by designing core pins at the size and draft angles identified in the sited previous work. The cast bosses were cut into nuts that could be used in the Ford Fastener Laboratory test-cell to measure clamp load when a thread forming fastener was driven into a cast nut. There were two sets of experiments run. First, a series of cast aluminum nuts were made reflecting the range of shape and size variations to be expected over the life cycle of a die casting die. Taptite thread forming fasteners, (a widely used thread forming fastener suitable for aluminum applications), were driven into the various cored, as-cast nuts at a constant input torque and resulting clamp loads were recorded continuously. The clamp load data was used to determine the range of clamp loads to be expected. The bolts were driven to failure. The clamp load corresponding to the target input of 18.5 Nm was recorded for each fastener. In a like fashion, a second set of experiments were run with cast magnesium nuts and ALtracs thread forming fasteners, (a widely used thread forming fastener suitable for magnesium applications). Again all clamp loads were recorded and analyzed similarly to the Taptites in aluminum cast nuts. Results from previous work performed on the same test cell for a Battelle project using standard M8 bolts into standard M8 nuts were included as a comparator for a standard bolt and nut application. The results for the thread forming fasteners in aluminum cast holes were well within industry expectations of +/- 30% for out of the box and robustness range testing. The results for the dry and lubed extreme conditions were only slightly higher than industry expectations at +/- 35.6%. However, when compared to the actual Battelle results (+/- 40%) for a standard bolt and nut the tread forming fasteners performed slightly better. The results for the thread forming fasteners in magnesium cast holes were all well within industry expectations of +/- 30% for all three conditions. The robustness range (.05mm larger and smaller holes than the expected wear pattern of a die casting die at full life cycle) results also fell within the industry expectations for standard threaded fasteners. These results were very encouraging. It was concluded that this work showed that clamp load variation with thread forming fasteners is consistent with industry expectations for standard steel bolts and nuts at +/- 30%. There does not appear to be any significant increase in clamp load variation due to the application of thread forming fasteners in as-cast holes of aluminum or magnesium over the effective life of a die casting mold. The fully implemented potential benefit of thread forming fasteners in as-cast holes of aluminum and magnesium is estimated to be 6 trillion Btu per year for North America. Economic benefit is estimated to be nearly $800 million per year. Environmental benefits and quality improvements will also result from full implementation of this technology.

36 MATERIALS SCIENCE↗