Mr. Chairman and distinguished members of the committee, thank you for inviting me to testify today. I am Paul Robinson, director of Sandia National Laboratories. Sandia is managed and operated for the U.S. Department of Energy by a subsidiary of Lockheed Martin Corporation. We are, first and foremost, a national security laboratory. Throughout the history of the nuclear weapons program, we have been responsible for the nonnuclear engineering development of U.S. nuclear weapons; and today we continue to share responsibility for the stewardship of all U.S. nuclear weapons in the stockpile. Our mission comprises the design, certification, and assessment of the nonnuclear components and subsystems of nuclear weapons; safety, security, reliability, and use-control; issues associated with the production and dismantlement of nuclear weapons; surveillance and support of weapons in stockpile; and substantial work in nuclear intelligence, nonproliferation, and treaty verification technologies.

In this statement, I will discuss Sandia’s Stockpile Stewardship Program, with emphasis on the Stockpile Life Extension Program, and describe how DOE’s Accelerated Strategic Computing Initiative (ASCI) and Advanced Design and Production Technologies (ADaPT) initiative support that responsibility. In addition, Appendix B describes Sandia’s international security activities in support of U.S. nonproliferation objectives. And I would also like to direct your attention to Appendix A, which lists an impressive catalog of accomplishments in Sandia’s national security programs during the last year.

Sandia National Laboratories and its sister laboratories (Los Alamos and Lawrence Livermore) perform a central role in DOE’s nuclear weapons Stockpile Stewardship Program. It is our job to assure confidence in stockpile safety and reliability and help DOE create an efficient production complex for the smaller stockpile mandated by the START initiatives.

When our laboratories stepped up to the challenge of stockpile stewardship without nuclear testing, we outlined our needs for new science-based facilities and capabilities to enable us to better address the challenge. During the Cold War, in addition to a vigorous program of nuclear testing, a continuous stream of weapon development programs kept our nuclear weapon and system design skills sharp. However, in the future, the engineers who must perform stockpile support and stewardship responsibilities will not have had original weapon system design experience. We are moving from experience-based stockpile management to science-based stockpile management. This shift makes the scientific and advanced engineering capabilities of the laboratories more important than ever.

For the past four years, I and my colleagues at the other Defense Programs laboratories and in the Department of Energy have communicated our requirements for stockpile stewardship in the post–Cold War era and have urged your support of the investments necessary to achieve this vision. We appreciate the support that this committee and others in Congress have shown for that plan. Today I want to describe to you how your investment in science-based stockpile stewardship has already begun to produce significant results and to urge your strong support of the budgets which the Administration has requested. This committee has been unwavering in its support of the nation’s nuclear weapons effort, and we hope that the Administration’s strong support for the FY99 budget will be reflected similarly by strong, bipartisan approval in the Congress.

Stockpile stewardship is the part of the DOE Defense Programs laboratories’ mission that maintains and ensures the safety, security, and reliability of the enduring nuclear weapons stockpile. We are developing new capabilities in the Stockpile Stewardship program. As part of this effort, we are strengthening the engineering technology base that supports component and subsystem and system design and production now and for the future.

Special facilities are either under construction or being assessed for aboveground simulation of nuclear-weapon physics and effects, including the Dual-Axis Radiographic Hydrotest facility (DARHT), Advanced Hydrotest Facility (AHF), the National Ignition Facility (NIF), and the X-1 Advanced Radiation Source. These facilities will provide a package of tools that will help replace underground nuclear testing with laboratory experiments.

Sandia’s Z Accelerator has met milestones for temperature and energy which will support the technology for the X-1 Advanced Radiation Source. During the past year, we attained spectacular results: world-record x-ray energies and powers and high radiation temperatures that permit this machine to be immediately useful for weapons physics experiments. Weapons scientists from Los Alamos and Lawrence Livermore National Laboratories are using the Z Accelerator for important stockpile stewardship work today.

DOE’s Accelerated Strategic Computing Initiative (ASCI) and Advanced Design and Production Technologies (ADaPT) initiatives are the other major components of Sandia’s science-based stockpile stewardship. ASCI is developing a new generation of computer simulation tools that are essential for our goal of simulation-based life-cycle engineering. ADaPT will provide a modern design and production capability to repair, requalify, or rebuild warhead components of nuclear weapons. Both programs will help us meet the challenges of enhanced surveillance and stockpile life extension in the present and future stockpile.

The objective of the Accelerated Strategic Computing Initiative (ASCI) is to generate advances in computational science and technology to hasten the shift from test-based methods of development and stewardship to computation-based methods.

ASCI builds on DOE’s leadership in new computing methods and on industry’s development of new kinds of supercomputers. This initiative seeks to develop computers capable of hundreds of trillions of operations per second (hundreds of teraflops). It will also develop a new generation of full-physics, three-dimensional computer simulation tools that are needed for our goal of simulation-driven life-cycle engineering that supports our stockpile stewardship responsibilities. These tools are being developed at the three DOE Defense Programs laboratories in collaboration with U.S. research universities and computer firms.

At Sandia, ASCI consists of programs in applications, problem-solving environments, and high-end computing. The applications program develops high-performance, full-system, full-physics predictive codes to support weapons performance assessments, refurbishment analyses, accident analyses, and certification. The problem-solving environments program is creating a computational infrastructure and operating environment that makes ASCI computational capabilities easily accessible and usable. The high-end computing program supports the development and acquisition of the more powerful high-end supercomputing capability required by the ASCI applications.

 Consistent with modern industry practice, Sandia designers increasingly depend on virtual modeling and simulation, spending more time computing and less time and fewer resources on physical prototyping and expensive performance testing. Our needs range from integrating, accessing, and preserving existing information (such as war-reserve parts data and the documentation of weapons structures) to creating, disseminating, and assessing new information, including predictions of age-related material degradation and complex accident scenario analysis.

The first major milestone of the Accelerated Strategic Computing Initiative was the development of the world’s first teraflop computer by Sandia and Intel, which established new records for computational speed last year. That machine is now installed at Sandia, where it is routinely used in safety, aging, and nuclear performance studies for real stockpile problems.

The goal of DOE’s ADaPT program is to furnish modern design and production tools capable of providing high-quality product to support a robust nuclear deterrent in a cost-effective and environmentally responsive way.

For a variety of security, business, or technical reasons, it is impossible to rely on industry for all the components required for nuclear weapons. This is particularly true for components that are produced in low quantities and are unique to nuclear weapons. Consequently, DOE must retain a viable in-house manufacturing capability.

It should be emphasized that the nuclear weapons program requires an intimate working relationship between the laboratories, where most of the technology is developed, and the DOE production plants, where component manufacturing takes place. Sandia designs or specifies nearly all of the nonnuclear components of nuclear warheads. Consequently, we work closely with the production engineers at Allied Signal, Federal Manufacturing and Technologies, Kansas City, who are responsible for manufacturing most of these components. We also collaborate with engineers at the Pantex plant in Amarillo, where warheads and bombs are assembled or disassembled. In addition, we produce some components in-house at Sandia, an assignment we received as a result of plant closures in the DOE complex. These local operations serve as a convenient test bed for the technologies and methods being developed under ADaPT.

The technologies and processes developed and deployed by ADaPT will allow us to design, develop, produce, and certify the high-quality replacement parts and components needed for the Stockpile Life Extension Program in an efficient and responsive manner. The initiative will help us exploit modern approaches to product design and production, which can also offer advantages to US industrial firms. The development and use of these advanced methods will also help us attract and retain the first-rate design and production engineers that we need to support the nuclear weapons stockpile.

One of the major long-term challenges we face is how to ensure the reliability of an aging stockpile in the absence of testing and new weapon development programs. We surveil and assess the stockpile to ensure that nuclear weapons continue to be reliable and safe and that they are upgraded as necessary to maintain their capabilities until system retirement. Unfortunately, we do not possess sufficient data on how reliability declines as systems get older than about 20 years. However, it will soon be our daunting task to ensure that systems remain reliable and safe for decades beyond their designed service lives.

Sandia is addressing these concerns through an Enhanced Surveillance Program, which involves fundamental research in materials aging, the study of the effects of aging in components and subsystems, and the development of computational tools to model and predict the effects of aging without resorting to destructive testing from the increasingly limited stockpile base.

The Enhanced Surveillance Program is proceeding with several projects. A Predictive Materials Stewardship Science Project is developing a calculational capability to quantify the performance and reliability of stockpile materials throughout their service lives. Materials used in the stockpile are scored based on their susceptibility to aging degradation. Materials of concern are prioritized by their potential impact on the safety, performance, and reliability of weapons.

The Enhanced Component Surveillance Project focuses on the impact of aging in components and subsystems. Stockpile components are prioritized to reflect our judgment of the risk of failure. We direct our efforts toward widely used components with identified aging concerns, such as electromechanical, electronic, and explosive subassemblies.

The Future Surveillance Tools Project focuses on the development of intelligent sensors, advanced communication technologies, and automated approaches to enabling more precise and accurate stockpile surveillance across a broader range of parameters, at lower cost, and with reduced human interaction.

The DOE Office of Defense Programs established the Stockpile Life Extension Program to maintain the safety, reliability, and performance of the U.S. nuclear deterrent without the complete systematic replacement of aging warheads, but by replacing certain warhead components. This program will define requirements for the future nuclear weapons complex, providing the detailed information necessary to develop an overarching stockpile support strategy.

The Stockpile Life Extension Program will permit us to schedule routine limited-life component exchanges and systematic upgrades in related subsystems and components in need of replacement. While primarily driven by the need to replace limited-life components, it will also upgrade the technological currency of components and help maintain a production workload free from peaks and valleys. We still have some way to go in planning the Stockpile Life Extension Program for a consistent and predictable workload that meets military requirements, but we are making progress.

The Stockpile Life Extension Program provides a planning process for evaluating components in every type of weapon in the nuclear weapons stockpile by focusing on each component’s contribution to reliability, performance, and safety over the long term. The program places particular emphasis on components whose degradation might cause a reduction in weapon performance or safety. It provides the planning factors for evaluating weapon refurbishment actions in light of such concerns.

Sandia’s stockpile stewardship strategy will develop or acquire design and production technologies to reduce the cost of performing required life-extension activities under the Stockpile Life Extension Program. The Accelerated Strategic Computing Initiative (ASCI) and the Advanced Design and Production Technologies (ADaPT) program are key to enabling us to achieve this goal.

I would like to use a real-world example to illustrate how DOE’s Accelerated Strategic Computing Initiative (ASCI) and Advanced Design and Production Technologies (ADaPT) initiative are already helping us improve the performance of our stockpile responsibilities.

Neutron generators are critical components of U.S. nuclear weapons. It is our job to assure that they will operate reliably at any time during a life span of many years and in extreme and possibly hostile environments. Sandia has designed neutron generators for decades, and in the last few years we have also assumed responsibility for manufacturing them for DOE. Neutron generators pose formidable engineering challenges, but DOE’s ASCI and ADaPT initiatives have given us powerful new tools for performing this job.

A few years ago, we undertook to design a replacement neutron generator for the W76 nuclear warhead on the Mark 4 reentry body of the Navy’s Trident I system. There were several compelling reasons for doing so, including the need to increase the component’s design margins, simplify its manufacturability, augment its resistance to new profiles of hostile environments, and increase its life span.

Neutron generators present unique challenges for design engineers. The physics of these devices are complex: they function as miniature linear accelerators, to produce deuterium-tritium reactions in order to generate neutrons. Consequently, their design parameters are sensitive to minute variations. They must be designed for ruggedness against severe environments such as acceleration, vibration, high voltage, radiation, and mechanical impulse. In the past, we have always relied on an iterative design process involving numerous physical tests and whatever modeling tools were practical at the time.

Neutron generators also present tough challenges for production engineers. These devices necessarily contain exotic materials that require special fabrication processes, and they must be manufactured for high reliability over many years. Processes such as brazing, welding, plating, metallizing, material deposition, and encapsulation must be performed to very high DOE quality-control standards.

Advances in computing hardware and software under ASCI have given us great leverage in the engineering design phase. Our designers have used a new integrated computer model that simulates the performance of the ferroelectric power supply in the neutron tube, including the combined pressure and electric fields. This model, which runs on our teraflop parallel computer, permits designers to visualize the electrical and physical performance of the device while it is still "on the drawing board," so to speak. That "drawing board" is a CAD (computer-aided design) model in today’s parlance, and the design engineer can change the model and re-visualize its performance on computer many times before committing to a physical prototype.

ASCI's braze furnace simulations are allowing us to set manufacturing parameters without difficult and expensive manufacturing process engineering development. ASCI also has allowed us to model critical encapsulation processes for manufacturing the generator.

Computer codes developed under ASCI will make it possible, for the first time in the history of the program, to certify (without underground testing) a neutron generator for resistance to hostile radiation effects, which is required before the device can be accepted as a war-reserve component for the stockpile. Relying on data from past tests and simulations, physicists exploited the new capabilities of the ASCI program to develop large, adjoint codes that simulate three-dimensional radiation transport and mechanical response. These codes, which run on our ASCI teraflop computer, have been validated against experiments performed on aboveground simulators. Simulations have identified failure probabilities at some locations in the design, and designers are able to correct those problems.

DOE’s Advanced Design and Production Technologies (ADaPT) Initiative and Advanced Manufacturing (AM) Program have also provided useful new capabilities that advance the neutron generator program. The ADaPT approach is to create an integrated product and process design (IPPD) environment with modern engineering tools. This environment will be achieved through the development or acquisition of advanced engineering design tools, advanced processes for component fabrication, and an enterprise integration effort that employs information technologies to control and manage product realization processes.

ADaPT’s process development program funds research and development of several process technologies, such as laser marking and welding, low-temperature brazing, and screen fabrication. This portfolio of projects is closely associated with ASCI-funded product realization technology projects in robotics, modeling and simulation, materials development, and control systems.

Under ADaPT, we have used computer design tools to evaluate the ion optics of the neutron tube, which is the key component of the neutron generator. This code reduced the time for performing the simulation from about three weeks to one hour. As a result, engineers were able to evaluate over 200 design variations within two months to optimize the design. In the past, only two or three variations could be evaluated prior to initial prototype build. Consequently, we were able to create a much better design and reduce the need for iterative prototyping.

We have also developed advanced expert systems that are being used to capture much of the "know-how" that our design and manufacturing engineers have developed over the past decades. These knowledge bases are now being tested for use in the evaluation of new designs and manufacturing processes. These software tools will allow future engineers to "consult" easily with the experts who designed and developed the weapons components now in the stockpile.

These efforts have directly contributed to the neutron generator design and production program in several important ways. Robotic casting has permitted us to directly fabricate neutron tube bodies without assembled joints. We have substituted low-temperature, integral metal-ceramic bonding for high-temperature bonding. We have developed improved manufacturing processes for piezoelectric materials as well as new processes that will help avoid high-voltage breakdown in the tube. An intelligence-based, automated brazing inspection technique is in development, as well as an inspection system for sprays and coatings.

The enterprise integration arm of ADaPT has furnished modern information tools to improve production management and operations. We use a commercial manufacturing information data base system for production management, planning, and inventory control. A secure network linking the production facility to suppliers has cut procurement time for parts and materials. We have also developed software to simulate the integrated production process and automate production scripting.

Sandia’s research organization supports the ADaPT program with investigations into materials and processing problems that require long-term solutions. These efforts include laser processes for net shaping of structures and sealing of weld joints, new chemical processes for ferroelectric powders to address the supply problem for this material, multimedia production scripting, ion beam research, and other projects.

The neutron generator recertification program for the W76/Mark 4 has benefited in concrete ways from DOE’s investment in supercomputing and advanced design and production technologies. Our goal is for Sandia’s production facility to begin delivering the new MC4380 neutron generators in October 1999. The ASCI and ADaPT initiatives, together with supporting research activities, have provided outstanding capabilities to perform this stockpile responsibility with greater confidence.

Recent investments in science-based stockpile stewardship, including the Accelerated Strategic Computing and the Advanced Design and Production Technologies initiatives, have already begun to show results in our mission applications. Our competencies for supporting the nuclear weapons requirements of the nation during a new era of reduced development activity are getting stronger. The evidence for this conclusion is visible in programmatic work today.

I commend and thank this committee, and others in Congress, for supporting this vision. I am becoming more confident that your investment in stockpile stewardship will keep the DOE Defense Programs laboratories strong and able to meet their mission of maintaining high confidence in the nuclear stockpile's safety, security, and reliability.

at Sandia National Laboratories 

at Sandia National Laboratories 

Laboratory capabilities in stockpile stewardship provide the technology base for supporting U.S. objectives in arms control and nonproliferation. This appendix reports on how Sandia employs its stockpile stewardship competencies to support international security programs that reduce the threat of proliferation of nuclear materials and expertise.

Sandia designs and implements physical protection systems for nuclear facilities and materials. As directed by DOE, we provide technical consultation on the design, evaluation, and implementation of physical protection systems for foreign nuclear facilities and materials, with the goal of reducing vulnerability to theft or diversion. These services are performed pursuant to statues and international agreements. Almost all of the projects described in this Appendix are carried out in partnership with the other Department of Energy national laboratories. These efforts represent some of the most successful joint programs of the laboratories. The descriptions below highlight Sandia’s contributions, but each of the laboratories are carrying out important assignments in a well-coordinated, cooperative effort.

As the United States and Russia reduce the large arsenals of the Cold War to the smaller stockpiles of the post–Cold War era, both the U.S. Department of Energy and the Russian nuclear complex will be responsible for safeguarding a growing inventory of nuclear-weapon materials and components. These materials present a great risk for nuclear weapons proliferation and a substantial risk for accidents and environmental contamination.

A new relationship is developing between our DOE national laboratories and their Russian counterparts. While both sides carefully protect classified and sensitive information, Sandia has been engaged for more than five years with the important Russian nuclear design laboratories—Arzamas-16, Chelyabinsk-70, and the Institute of Automatics—as well as other Russian institutes such as the Kurchatov Institute and Eleron. Our goals in these engagements are to support U.S. national security by understanding and influencing these core organizations in the Russian nuclear weapons complex and to jointly develop technologies and approaches that will provide both sides with protection, accounting, and transparency. 

Sandia’s International Security Program has worked toward these goals by supporting numerous projects in the former Soviet Union on protection and security of nuclear materials and facilities. Cooperative interactions in these matters help achieve other desirable objectives, such as open dismantlement of all types of weapons of mass destruction, cooperation on nonproliferation activities, conversion of defense-oriented capabilities to civilian enterprises, and greater access by Western companies and laboratories to world-class science and technology within the former Soviet Union.

One of the key goals of the International Security Program at Sandia is to achieve worldwide protection and control of nuclear materials and weapons. A major step toward realizing this goal is our work with the former Soviet Union on Material Protection, Control, and Accounting (MPC&A). Other projects are also underway that contribute to this goal, including Cooperative Measures, Storage Monitoring, and Initiatives for Proliferation Prevention Program (IPP).

The Material Protection, Control, and Accounting program was established in 1994 to encourage the DOE Defense Programs laboratories to cooperate with the former Soviet nuclear institutes to protect nuclear materials that could be used to make nuclear weapons. The program includes American, former Soviet, and international partners. Its goal is to reduce the threat of nuclear proliferation and nuclear terrorism by rapidly improving material protection, control, and accounting at all facilities that contain weapon-usable nuclear materials. Activities under this program include developing and installing improved protection systems, identifying and implementing training and procedures, and strengthening site-level and national standards and regulations.

Today, Sandia is involved in security projects at 53 sites in the former Soviet Union. We work with nuclear materials specialists there to help them prevent theft of nuclear materials at installations and during transport. Sandians are also involved in a number of projects related to regulatory development and training for materials protection, control, and accounting. Specific improvements over the last year include the following examples.

Sandia has taken a lead role in developing cooperative relationships with the important organizations in the Russian nuclear complex. Several of Sandia’s senior executives have regular interactions with their Russian counterparts. These discussions are carefully planned and conducted. They are followed up by our staffs in order to promote understanding of the issues of concern on both sides and to develop high-level support and leadership for our jointly developed technical programs. Our interactions and contracts with Russian institutes are carefully monitored and carried out under DOE guidance.

Lab-to-lab contracts for research in safety technologies have been one of our primary activities with Russia. These contracts have provided new information for Sandia’s safety program in the areas of safe containers for transportation and storage, safety assessment techniques, accident consequence data and modeling, and studies of the safety characteristics of materials.

I believe both sides have benefited from these safety and security exchanges, particularly in meetings where we share information about accidents, accident environments, and accident response capabilities. We particularly value the participation of the Russian Ministry of Defense and our own Department of Defense in these technical exchanges.

Sandia has developed monitoring technologies for DOE storage operations to reduce cost, manpower, and radiation exposure of personnel while enhancing protection and accounting. As an outgrowth of this work, we have recently begun joint development with Russia’s Arzamas-16 site of such storage monitoring technologies with the added aspect of transparency.

The potential for high transparency in storage monitoring is apparent in our joint demonstration on the Worldwide Web (http://132.175.176.83/index.htm). This demonstration features live data from a storage container in the formerly secret city of Arzamas-16 and comparable data from Sandia’s sites in New Mexico and California. While no nuclear materials are involved in the demonstrations, the concept of live data accessible to anyone in the world transmitted from deep within the Russian nuclear complex would have been unimaginable only a few years ago. Such breakthroughs will be required for the new relationship between the United States and Russia to evolve toward a safe and secure future.

The Initiatives for Proliferation Prevention program provides a mechanism for scientists and engineers who have been supporting research and development on weapons of mass destruction in the newly independent states of the former Soviet Union to build careers in the burgeoning Russian civilian workplace. The program makes use of capabilities in DOE’s national laboratories and makes new technologies available for commercialization by U.S. and Russian industry. Sandia has been involved in 125 projects totaling nearly $19 million with over 40 participating institutions in the former Soviet Union. Fifty-one projects have been completed, 27 are active, and several proposals are awaiting final approval. Included among these projects are several prospective business enterprises involving nuclear-complex personnel.

Sandia participates in international efforts for securing nuclear material in nations other than the former Soviet Union. These activities are classified in three categories: 1) physical protection bilateral exchanges, which focus on the protection of U.S.-origin nuclear materials in the possession of other countries; 2) physical protection training presented to an international audience; and 3) cooperation with the International Atomic Energy Agency on the International Physical Protection Advisory Services.

Sandia works with countries with which the United States has signed bilateral Agreements for Peaceful Uses of Atomic Energy. (Some of these agreements are trilateral, with the International Atomic Energy Agency [IAEA] as the third participant.) Such countries use U.S.-origin nuclear materials. Sandia’s work within the framework of U.S. bilateral agreements fulfills statutory obligations.

Bilateral projects consist of discussions to promote technical exchanges and cooperation in matters relating to physical protection of nuclear materials. Information exchanges include legal and regulatory issues, technology, design basis threats, and transportation concerns.

Since 1974, experts from Sandia have participated with U.S. teams that made 80 trips to 48 foreign countries to participate in cooperative work. The teams are led by the Department of Energy and include participation by the Nuclear Regulatory Commission and Defense Special Weapons Agency.

Bilateral partnerships are finding renewed emphasis in the 1990s as a consequence of increasing concerns over nuclear nonproliferation and nuclear smuggling.

When the United States passed the Nuclear Nonproliferation Act of 1978, it obligated the Department of Energy to share physical protection technology with the member states of the International Atomic Energy Agency (IAEA). Sandia conducts the International Training Course on Physical Protection of Nuclear Facilities and Materials on behalf of DOE under the general auspices of the IAEA.

The course, conducted in the United States, trains participants in a systematic methodology for designing and analyzing physical protection systems to protect nuclear facilities and materials from sabotage and theft. Since 1978, Sandia has presented this course fourteen times to over 400 participants from 57 countries.

Sandia conducts additional training directed at the needs of specific countries or regions as a supplement to the International Training Course. In addition to providing training on basic physical protection methodology, specialized topics on vulnerability assessment, response force procedures, intrusion detection sensor installation, and transportation security have been taught. Country-specific physical protection training has been conducted in South Korea, Brazil, Colombia, Belarus, Kazakhstan, Lithuania, Russia, and Ukraine. Regionally, physical protection training has been conducted in the Czech Republic for Russian-speaking participants from Eastern European countries. This year, plans are underway to conduct a regional training course in Argentina for Spanish-speaking participants from Latin America and a country-specific course in China. 

In cooperation with the U.S. Department of Energy and the International Atomic Energy Agency, Sandia has supported a newly formed International Physical Protection Advisory Service (IPPAS). IPPAS is a voluntary service requested by IAEA member states for an independent review of their physical protection measures. The review can be all-inclusive (reviewing national nuclear physical protection regulations, local physical protection measures, and specific facility practices) or specific to certain physical protection issues.

Sandia supports this activity by contributing technical consultants to the international teams of reviewers. The product of an IPPAS mission is a report of findings, recommendations, and good practices. If warranted, an IPPAS mission may be followed up with return visits, training and equipment to fix noted deficiencies at nuclear facilities.