Advances in Cryogenic Engineering

Sam Collins was a perfect gentleman in the traditional sense. He worked without fanfare and had no time for image building. He wanted his work to speak for itself and his papers were modest descriptions of his major accomplishments. Many of his original ideas were never published; nevertheless they have served as the basis for the achievements of others.

The SSC is a high luminosity pp collider designed to achieve 40 TeV in the center of mass. Depending on the final magnetic field chosen the main ring will be between 90 and 165 km in circumference. Construction of the SSC has been recommended to the DOE by HEPAP* for completion in the early 1990’s. The Universities Research Association has been designated by the DOE to do R/D and prepare a design proposal, construction plan and cost estimate. Model magnets are being tested and a field level will be chosen before October 1985. A design proposal will be submitted in April 1986.

A design is presented for a dipole magnet suitable for the proposed SSC facility. Test results are given for model magnets of this design 1 m long and 4.5 m long. Flattened wedge-shaped cables (“keystoned”) are used in a graded, two-layer “cos θ” configuration with three wedges to provide sufficient field uniformity and mechanical rigidity. Stainless steel collars 15 mm in radial depth, fastened with rectangular keys, provide structural support, and there is a “cold” iron flux return. The outer-layer cable has 30 strands of 0.648 mm diameter NbTi multifilamentary wire with Cu/S.C. = 1.8, and the inner has 23 strands of 0.808 mm diameter wire with Cu/S.C. = 1.3. Performance data is given including training behavior, winding stresses, collar deformation, and field uniformity.

Magnet Design B for the Superconducting Super Collider (SSC) consists of a 5 cm diameter collared coil assembly 12 m long with concentric aluminum thermal shields at 10 K and 80 K, a G-10 post type support system and a minimal iron vacuum vessel located at a large radius from the coil. In order to determine the behavior of such a magnet under both direct current and quenching conditions, a 6 m model was built using Tevatron tooling to produce a 7.6 cm diameter coil. The dc operation demonstrated that the post type suspension has acceptable rigidity. Distortions in the aluminum thermal shield during quench resulted from stresses in the material below the yield values. Temperature increases in the thermal shield due to eddy currents were larger than those calculated using simple assumptions, demonstrating the value of using a model to verify eddy current behavior in complex situations.

In order to proceed with the design of magnets for the superferric 3 tesla version of the Superconducting Super Collider (SSC), it was necessary to simulate quenches by the use of computer programs. For this simulation, we used two different programs, a modified version of ‘QUENCH’, and ‘SSC*’, a program developed specifically for this magnet design.

The design of the 4.5 K primary cooling system and higher temperature shield cooling systems for the SSC are described. Typical flow diagrams for the magnet piping systems are presented. Estimated heat loads are given. The systems have been designed to accomodate the great distances, 90 km and up, over which the load will be distributed. Provision has been made for cooldown, warmup, quench recovery and magnet replacement, as well as for steady-state operation.

A design has been developed for an SSC 6T option dipole magnet cryostat. The design criteria that defines the basic parameters and performance requirements are discussed. Details of the single phase assembly, suspension, insulation, thermal shields, vacuum vessels and interconnections are presented. Results of the experimental program in support of the design effort are discussed.

A support member for superconducting magnets and other cryogenic devices has been designed, fabricated and structurally and thermally evaluated. The member is a cylindrical post constructed with fiber reinforced plastic (FRP) tubing and having metallic heat intercepts and end connections. All FRP to metal connections are made by mechanical shrink fitting and do not employ adhesives or fasteners. The post can operate in tension, compression and flexure or in combinations of these loads. The details of the design and construction are enumerated. Structural performance has been measured in tension and compression at 80 and 300 K and in flexure at 300 K. Creep effects on the shrink fit joint reliability are being evaluated. Thermal performance has been measured for a post with ends at 4.5 and 300 K and with heat intercepts at 10 and 80 K. The measured performance has been compared with the analytical predictions. Full scale, working, prototype posts have been successfully utilized in several model cryostats for the Superconducting Super Collider dipole magnet development program.

One of the interesting problems associated with building long magnets for the SSC (Superconducting Super Collider) is predicting and controlling the dynamic response of the cryostat tubes during cooldown. Thermal bowing occurs in any of these tubes that are asymmetric in shape or which are not cooled uniformly. Understanding the bowing behavior is important for two reasons: First, one needs to know the magnitude of the induced displacements so that potential interferences in the entire magnet assembly can be located. Second, the bowing phenomenon introduces structural loads on the supports which need to be folded into the design of those supports. It is desirable, due to cost and time constraints, to develop an analytical model which accurately predicts loads and displacements rather than relying on a physical model of each candidate cryostat tube design. This report describes a procedure and an analytical model to predict this dynamic behavior on the thermal radiation shield for Fermilab’s proposed SSC magnet design. The results are compared with test data obtained on a physical model fabricated and tested in an effort to verify the analytical approach.

The superferric magnet option for the SSC requires approximately 160 km of magnets. In order to reduce the number of field connections, it is proposed to use 150 metre long magnet assemblies. Each of these assemblies contains four dipole magnets of 35 metre length, one quadrupole magnet, trim coils and a piping and wire assembly which provides the capability of taking up 45 centimeters of thermal contraction. The cryostat contains the magnet vessel, a couple of heat intercepting shields, superinsulation and pipes carrying refrigerant to and from the refrigerators, located at approximately 6.5 km intervals. The magnet vessel contains the iron laminations, beam tubes, helium channels, busbars and magnets. Weight of the magnet vessel assembly is of the order of 240 kg per metre of length. A large number of problems need to be solved in the cryostat design. Because of cost, available space for the cryostat is limited. Cooldown of the system requires flow channels of large cross- section. Removal of synchrotron radiation during steady state requires high flowrates in the area of the beam tube, where space is limited. Fast warmup and cooldown in local areas is required to repair and/or replace faulty magnets. Non uniform cooling will result in bowing. Beam tube vacuum requires special consideration when pumping kilometers of small diameter tubes. Aligning magnets after installation in the tunnel requires knowledge of the location of the magnet assembly in a cryostat.

Superconducting Magnetic Energy Storage (SMES) has potential as a viable technology for use in electric utility load leveling. The advantage of SMES over other energy storage technologies is its high net roundtrip energy efficiency. This paper reports the major features and costs of a 5000 MWh SMES plant design.

A new two-layer SMES solenoid design is presented as a possible replacement for earlier low aspect ratio single-layer Wisconsin SMES designs. The two-layer low aspect ratio design maintains most of the advantages and simplicities of the single layer design with several added benefits: lower current, smaller conductors at 200 kA which replace 800 kA conductors; more accessible and easier mounted conductors in two vertical layers, one inside and the other outside a solenoidal axial structure wall; improved conductor mechanical support due to the tapering together of the axial structure at the top and bottom to bring the end turns of inner and outer layers closer together; and an improved flexibility for force and field optimization for a thicker wall two-layer solenoid. Advantages for both double and single-layer coils are: compatible thermal-mechanical mounting on bedrock trench walls to contain expanding forces; rippled structure and conductor to reduce cooldown and magnetic loading stresses; open mounting in 1.8 K pool helium for good cooling; round conductor continuous assembly without joints; convenient design to short all turns in parallel for emergency discharge; and relatively low voltages for such large storage inductors. The key features of system stress analyses and system conductor stabilities for the new two-layer design will be described.

Superconducting Magnetic Energy Storage (SMES) is a potentially cost effective technology for electric utility load leveling. Design concepts and cost estimates of SMES plants capable of delivering 5000 MWh daily have been previously identified. An important feature of a large commercial plant is a system that will reliably shut down the magnet by thermally dissipating the stored energy in the event of an imminent or actual loss of superconductivity. To prevent damage to the coil during such a protective energy dump, the entire coil must be driven “normal”, i.e., resistive rather than superconducting, in a short period of time. This requires rapid removal of the liquid helium coolant surrounding the coil.This paper describes a simple system that has been developed to rapidly remove the liquid helium from the helium vessel which surrounds the coil. The system has only a small number of active components, no external helium storage, and is practical to reset and maintain.

Consideration of the environmental impact of SMES systems includes possible biological effects on humans as well as on other vertebrates, plants and bacteria; effects on the operation of essential life support systems such as cardiac pace makers; and effects on such equipment as watches, microprocessors, automobile and aircraft ignition systems and magnetic credit cards. Present knowledge of the biological effects of steady (DC) high and low intensity magnetic fields is reviewed, including synergetic effects of such fields in the presence of 60 Hz electric and magnetic fields; effects on cell growth, DNA synthesis, endocrine system rhythms, Ca++ efflux, bacterial motion and bird migration are considered briefly. If the environment outside the fenced-in SMES area is to be accessible to persons with cardiac pace makers the safe field level will have to be below 1.7 mT. Selecting an arbitrary safety factor of 1.7 giving a field of 1.0 mT one obtains an exclusion radius of 2.0 km for a presently considered 5500 MWh solenoidal storage system and 255 m for a proposed 20 MWh device.

Recent proposals for ignition tokamaks containing superconductors are reviewed. As the funding prospects for the U. S. fusion program have worsened, the proposed designs have been shrinking to smaller machines with less ambitious goals. The most recent proposal, the Tokamak Fusion Core Experiment (TFCX), was based on internally cooled cabled Nb₃Sn conductors for the options which used superconductors. Internally cooled conductors are particularly advantageous in their electrical insulating properties and in the similarity of their winding procedures to those of conventional copper coils. Epoxy impregnation is possible and is advantageous both structurally and electrically. The allowable current density for this type of conductor was shown to be larger than the current density for more conventional superconducting technology. The TFCX effort identified research and development needed in advance of TFCX or any other large ignition machine. These topics include the metal used for the conduit; nuclear effects on materials; properties of electrical and thermal insulators; extension of superconducting technology to the sizes of coils envisioned and to the field level envisioned; pulsed coil superconducting technology; joints and insulating breaks in conductors; heat removal or flow path length limitations; mechanical behavior of potted conductor bundles; instrumentation; and fault modes and various questions integrated with overall machine design.

The MIT 12 tesla coil was fabricated using a 486-strand bronze-matrix Nb₃Sn, JBK- 75 superalloy sheathed Internally Cooled Cabled Superconductor (ICCS). The rectangularly shaped ICCS conductor was wound into three double pancakes prior to a six-day reaction (4 days at 700 C followed by 2 days at 730 C). Prior to installation in the High Field Test Facility (HFTF) at LLNL the three subcoils were insulated and epoxy potted. With the HFTF providing a 9 T background field, the test coil was used to raise the central field up to 12 T. Measurements of critical current are reported for fields in the 11 to 12 T range for temperatures of 4.2, 5.2 and 7.5 K with supercritical as well as 1 atm, two-phase internal helium. Steady-state, inductively heated as well as quench induced operations are described.

The first bipolar operatinal test of an MJ-class pulsed superconducting magnet was successfully carried out in energy transfer experiments between a 3 M J magnet(load magnet) and a 4 MJ magnet(energy storage magnet) using a chopper-type energy transfer equipment. The test goal was to confirm the technical feasibility of operating an ohmic heating coil for a tokamak-type fusion reactor. A 3 MJ magnet was first charged to 4510 amperes in 1.5 seconds. It was discharged to 0 ampere in 1.5 seconds and at once was re-charged to −4510 amperes in 1.5 seconds. It was again discharged in 1.5 seconds. The maximum field and the stored energy were 5.4 T and 2.1 MJ, respectively. During this bipolar sweep test, the magnet was quite stable during charge. Instability phenomena were not observed when the direction of the current was reversed. We also found that there was no mechanical instability during operation from the results of dynamic strain measurement. Thus it can be concluded that there is no technical problem in the bipolar sweep of an MJ-class pulsed superconducting magnet.

The present work generalizes the results on testing a full-scale superconducting dipole magnet in force-circulating cooling mode. The dipole characteristics and description of the force-circulating test facility are presented. The tests were performed with the help of single and two-phase liquid helium at a temperature of 4–4.8 K. The quench current is the same as that of a short sample superconducting cable and exceeds 7.2 kA for the 6.1 T field induction in the magnet center. The results on measuring the dynamic characteristics and ac losses are also presented.

An iron-free superconducting quadrupole with a 140 mm coil inner diameter and a, field gradient of 70 T/m has been built and tested as a prototype magnet for use in low beta insertions of the TRISTAN Main Ring. The magnet is made from 27 strands of NbTi Rutherford type cable clamped with stainless steel collars. A maximum field gradient of 80 T/m with a peak field of 6.0 T in the coil was obtained after five quenches. Quench characteristics and field gradient homogeneity of this quadrupole were studied in a vertical cryostat. Mechanical disturbance energies in the coil were also obtained from measurements of the voltages across the coil.

The 3 m Φ × 5 m long × 1.5 T superconducting solenoid for the Fermi- lab Collider Detector has been installed at Fermilab and was tested in early 1985 with a dedicated refrigeration system. The refrigerator and 5.6−Mg magnet cold mass were cooled to 5 K in 210 hours. After testing at low currents, the magnet was charged to the design current of 5 kA in 5−MJ steps. During a 390 A/min charge a spontaneous quench occurred at 4.5 kA due to insufficient liquid helium flow. Three other quenches occurred during “slow” discharges which were nevertheless fast enough to cause high eddy current heating in the outer support cylinder. Quench behavior is well understood and the magnet is now quite reliable.

Fermilab has successfully tested an iron bound 3 tesla superconducting solenoid for the Tohoku Bubble Chamber, Thermal performance, magnetic field calculations, charging characteristics, and a special Ni30%Fe dump resistor are reported. Low heat leak is obtained by thermally intercepting the supports with boiloff gas.

In the High Field Laboratory for Superconducting Materials at Tohoku University, we have built up in the last four years three hybrid magnets capable of 30 T, HM-1, HM-2, and HM-3. Two of these magnets, HM-3 (32 mm bore, 20 T) and HM-2 (52 mm bore, 23 T) have been already made available for the development of superconducting materials for nuclear fusion reactors. The outer, superconducting, parts SM-3 and SM-2 of these magnets, wound with multifilamentary NbTi conductors, can both generate 8 T and have the room temperature bores of 220 and 360 mm, respectively. The fabrication of HM-1, the maximum field of which had been designed to be either 31 T (32 mm bore) or 29 T (52 mm bore), was finished last spring and we succeeded in generating 30.7 T on May 29, 1985. A 12-T superconducting magnet (SM-1) with a room temperature bore of 360 mm was used in HM-1. SM-1 used Ti-doped Nb₃Sn multifilamentary couductor. The design, construction, test and operation of these three superconducting magnets are reported in detail. A few examples of H_C2 and J_c measurements in HM-1 are also presented.

This superconducting magnet system has five components: a two-coil 11-T superconducting magnet, 150 A power supply, a metallic inner dewar, 80-L liquid helium outer (magnet) dewar and their auxiliarys. The magnet is 40 mm in inner diameter and 257 mm in length. It has a stored energy of 73 kJ at a maximum field of 11 T. The inner coil is made of Nb₃Sn/Cu conductor (Cu:SC ≃ 4:1) and the outer uses NbTi/Cu conductor (Cu:SC ≃ 2.3:1). (Both superconductors are supplied by two Chinese Institutes). Full field is achieved at 50J of the short sample current density. The inner dewar provides a sample volume with a diameter of 30 mm which can be varied in temperature from 4.2 K to 400 K. Samples can be changed when the magnet is operating. The loss rate of the outer dewar itself is 0.17 litres per hour. At 10 T, with the temperature of the inner dewar at 70 K, the loss rate is approximately 1.3 litres per hour.

A vapor-cooled lead system that used six pairs of leads in twelve separate dewars was built for the International Fusion Superconducting Magnet Test Facility (IFSMTF). Each of these leads was cooled by helium vapor from a reservoir separate from the pool-boiling or force-flow coil. Satisfactory acceptance tests were performed on the early production runs of the leads. Inconsistent and anomalously high heat losses were observed on the two pairs of lead assemblies used in the partial-array test. Subsequent tests on the leads and their associated dewars confirmed satisfactory performance of the lead and dewar and offered an explanation for the excessive losses during the partial-array test.

Japan Atomic Energy Research Institute (JAERl) has been developing a high-current vapor-cooled current lead with a nominal current more than 30 kA for used with the superconducting coil of the Fusion Experimental Reactor. As a development step, we have fabricated a number of current leads with a cable-in-conduit geometry and have tested them up to 20 kA. The measured heat leak at 4 K is about 1.0 W with helium flow rate of 0.05 g/s-kA, which is exactly the same as for the ideal self-cooling condition. In addition, stable operation at the nominal current is possible for more than 10 min with no helium flow. The measured thermal performances are in sufficiently good agreement with those calculated that we concluded that the heat exchange coefficient in current leads is very close to the theoretical value. This paper describes the development status and the test results obtained at operating currents up to 20 kA together with the comparison between the measured and calculated characteristics.

The heat flow through current leads is one of major heat losses in a superconducting magnet system. To reduce the heat flow, current leads have been optimized in a complex way by varying such quantities as conductor length, cross-sectional area, heat transfer coefficient and cooling perimeter. Therefore, this study is made to simplify the design procedure, and to explain the general characteristics of the current leads. A new combined parameter which takes turbulent flow into account is introduced in the present work to enable us to draw a useful design chart. This chart gives, to a wide variety of current leads, detailed information about the optimum design-viz. geometric dimensions, heat flow into liquid helium, and pressure drop of the cooling gas. Change of the cross-sectional area along the conductor may improve the current lead performance. The effects of this area change are examined in detail.

Usually mechanical switches that are built for use in superconducting circuits are driven in some way by a rod which is controlled at room temperature. In this paper, an alternative method to drive the electrodes of the switch is reported. In fact the new device is a superconducting relay that uses an antiseries connection of two superconducting air-core coils. The repulsing force of these relay coils enables the switch to be closed by applying a pressure to the electrodes. The off-state is effected by a set of springs which interrupt the electrodes when the coil current is switched off. We realized that this electro-magnetic method of producing large forces could be promising for driving a mechanical switch. The desired method was demonstrated by an experimental model. A switch-on resistance of 8 ß 10−8 Ω with a switch current of 3 kA and a contact force of 20 kN was measured.

A rotating coupling for a superconducting prototype generator was developed and tested. It connects the rotor to the helium liquefier and the external vacuum pump. One liquid and four gas helium streams are separated from each other by labyrinth seals. The vacuum connection is located around the centerline and sealed by a magnetic fluid seal. A special arrangement of the liquid helium duct provides self-acting regulation of the liquid helium level in the rotor.The rotating coupling was tested in a special test facility. It was operated at rotational speeds up to 3000 rpm at room temperature and for about 90 hours with liquid helium. Proper functioning of the automatic liquid level regulation system was established. All the welds and the magnetic fluid seal remained vacuum-tight during operation.In a separate long-term test a magnetic fluid seal was successfully operated for almost four years.

Acoustic emission (AE) diagnostic and monitoring techniques have proven powerful in identifying and understanding causes that affect the performance of superconducting magnets. This paper reviews our recent progress in this area, focusing in the applications to: 1) quench source identification, 2) disturbance energy quantification, 3) normal-zone detection, 4) strained Nb₃Sn wires, and 5) electric arc detection. Future directions of research needed to achieve better reliablity are briefly discussed.

Acoustic Emission (AE) techniques are used to study the structural integrity of large superconducting magnets, particularly of fusion-device scale. The Large Coil Task (LCT) at Oak Ridge National Laboratory is designed to test six large superconducting coils in a tokamak configuration. We have used the LCT to explore long-term trends in AE characteristics and to develop improved methods for data acquisition and processing. Thirty differential AE sensors were strategically mounted onto the General Dynamics LCT Coil. This coil and the JAERI coil were tested during the Partial Array Test in the summer of 1984. Since the General Dynamics coil has not failed at this time, it is not possible to determine finally which AE parameters will best characterize magnet performance. Presented here are spatial distributions of AE trends which were observed and can be expected for this style of superconducting coil.

The objective of acoustic emission measurement at Japan Atomic Energy Research Institute (JAERI) is an establishment of a general diagnostic method for superconducting magnet systems. Output of strain and displacement gages can not cover a whole system in monitoring premonitory phenomena of a magnet system s failure, because these sensors are mounted on points and therefore localized. Acoustic emissions can be transmitted to sensors through structural materials without electrical noise. Monitoring of acoustic emission will be one of the methods to predict a serious failure of magnet systems in a vacuum vessel. For this purpose, several sensors were installed on the Japanese LCT coil and the Test Module coil (TMC). Some of acoustic activity was similar as seen in these coils. The correlation between voltage spikes and acoustic events is excellent during single coil charging mode, but poorer during out of plane force mode. There are no indicative acoustical phenomena before a magnet quench or during normal zone generation. The conditioning of acoustic events and voltage spikes can be seen after any cooling down. The localization of electrical insulation damage with the acoustic emission technique is one of its most useful applications.

A disappointingly low withstand voltage capability was found during high-potential testing of an electrical system consisting of a large superconducting coil and the equipment connected as it was installed in the International Fusion Superconducting Magnet Test Facility (IFSMTF)— two superconducting buses, two vapor-cooled leads, and 120 sensor cables with ambient temperature and cryogenic vacuum feedthrough connectors. DC and transmission line techniques were unsuitable for finding the location of the dielectric breakdowns. An acoustic emission (AE) measurement system was developed with which to determine the location of breakdowns in large coils after installation in IFSMTF. Using triangulation with AE sensors, the system measures the difference in time-of-arrival of transient waveforms caused by the dc voltage discharge. The system was calibrated on a stainless steel surface representing the coil case, and its accuracy was found to be better than 5 cm. This paper describes both the acoustic emission measurements and the high-voltage testing system used concurrently. Also presented are the experimental results of a series of high-voltage tests that led to the determination of the exact location of one breakdown in the GD/C coil system of the Large Coil Task (LCT) and the results of the successful repair.

A technique combining acoustic emission (AE) and voltage measurement has been applied to a quench experiment conducted on a newly fabricated multi-filamentary 15.1 tesla (T) superconducting magnet system. The system consists of a wind and react Nb₃Sn insert and dry wound NbTi background solenoid. The magnet system was charged in (i) He I and (ii) pressurized He II. The magnet system experienced many premature quenches, which were preceded by AE and voltage spikes. Quenches for the Nb₃Sn insert were due to the debonding between the winding and the magnet form, while those for the NbTi magnet were due to conductor motion. The magnet system finally attained the short sample critical current, where neither AE burst nor voltage spike appeared prior to the quench: They were 485 A, 13.6 T, in He I and 540 A, 15.1 T, in He II.

An instrumentation system for real-time location of acoustic emission (AE) events and identification of sources has been developed for monitoring high-performance superconducting dipoles and quadrupoles used in high-energy accelerators. AE energy, event count, and time of occurrence are extracted in hard-wired circuitry; digitized, stored, and transferred to a PDP-11/23 microcomputer. The system uses first hit method of event detection, with a time resolution of 30 µs — a substantial improvement over currently available commercial AE equipment. The system can be used for up to six AE channels simultaneously.

The acoustic emission (AE) from FRP cryostats has been studied in order to establish a monitoring system for an FRP cryostat. The Kaiser effect has been found in FRP cryostats and hence using the AE method we can in principle estimate the stress history of any FRP cryostat. The condition of the coolant was also detected by AE techniques. Sudden decreases of internal pressure which lead to the rapid evaporation of coolant can be detected as sudden increases of AE activity. It is also found that the vacuum leak area caused by gaseous pressure could be located by the processing of AE sources.

Epoxy cracking in epoxy-impregnated superconducting magnets is caused by transverse shear stresses created by Lorentz forces. The dissipative energy of cracking can induce premature quenches in these magnets. Dissipative crack energies have been measured in a test coil; the maximum energy measured in the test coil was ~2 mJ.

Stress-induced dissipative energies were measured at 4.2 K for an epoxy resin and mica with a newly developed high-resolution (~0.1 mJ) energy transducer. The dissipative energies have also been correlated to acoustic emission (AE) signals. Correlation indicates that it is possible to use AE signals to quantify these dissipative energies. Epoxy resin and mica are important materials in the winding of high-field Nb₃Sn superconducting solenoids required for high-frequency NMR spectrometers.

Mechanical properties of internally-cooled cabled superconductor (ICCS) under cyclic loading were investigated. Conductor strand motion due to the inherent lack of internal strand support in an ICCS, coupled with the strain sensitivity of NbsSn, could lead to premature failure of the Nb₃Sn superconductor. A two-step approach was used to study this phenomenon.First, two small test coils were non-inductively wound with 27-strand NbTi ICCS of 64% and 32% void fractions respectively. The coils were subjected to cyclic Lorentz forces and monitored for strand motion using a voltage transient/acoustic emission technique. Motion- induced voltage transients were observed in the 64% test coil and no voltage transients were observed in the 32% test coil.The second step involved subjecting test coils wound with 58% and 36% void fraction 27-strand Nb₃Sn ICCS to cyclic Lorentz forces. After 100 cycles, the 58% test coil showed a 24% degradation in critical current with no further degradation; the 36% test coil showed no degradation after 750 cycles of testing.

An overview of large magnet systems which have been studied, constructed, or operated in the last 12 years is presented and shows a substantial advance in overall current density, stored energy, and magnet complexity. The preferable coolant mode for very large magnets is still a bath of helium I, but it is clear that other coolant modes are gaining acceptance. The data base for design using stability criteria dependent on transients has expanded to the point where the risk is often acceptable, compared to the lower current density, low risk, steady state stability criteria which launched large superconducting magnet technology. The limitation imposed by structure and protection on increasing overall current density in large magnets is discussed and a simple model is used to illustrate the extreme requirements imposed on a winding without direct helium contact. The latter implies that a significant technological step is required before conduction cooling or indirect cooling will be used in the large magnets envisioned for the future and that helium contact with the conductor will remain the key ingredient for risk reduction in large magnet design.

Examination of the criteria governing the design of superconducting adiabatic windings shows that the extrapolation of the design of a magnet storing 6 MJ of energy to higher energies can lead to adiabatic windings of about the same size as their cryostable equivalents. The criteria considered are i) stability against the energy released by epoxy cracking, ii) quench hot-spot temperature, iii) quench voltage, iv) self-field instability and v) temperature rise due to steady heat dissipation within the winding. The dimensionality of the quench in an adiabatic winding is seen to affect the extrapolation. A winding in which the quench propagates for a significant time in three mutually perpendicular directions leads to the highest winding current densities, but to less tolerance for distributed internal dissipation, Extrapolation of operating current may be limited by self-field instability to lower values than are usual in large systems.

Several different formulas have been proposed for estimating quench energies of potted magnets.1–3 These formulas, although based on plausible approximations, disagree with one another by factors of 2 to 6. By the use of maximum and minimum principles for the heat balance equation, it has been possible to determine whether each estimate is an upper or lower bound.

The combined stored energy and current density obtainable with pool boiling magnets is constrained by a maximum stress limit, a quench adiabatic temperature rise limit, and a cryostable heat flux limit. The temperature rise and cryostability constraints can be combined to yield a relationship between practical maximum current densities and stored energies. The resulting curve, which bounds the realm of applicability for pool boiling design, is presented and compared with the requirements of several magnet applications. The ratio of maximum stored energy to current density is proportional to the square of the design heat flux. This, in turn, is determined by the choice of stability criterion and the available heat transfer within the winding. As the stability criterion becomes more liberal, the disturbance energies for which the conductor will recover become smaller. This performance/risk tradeoff is discussed in relation to the “training” phenomenon of pool boiling magnets and to the increase in resistivity of copper in magnets subjected to neutron radiation. Finally, it is known that cooling channel geometry and the amount of vapor present within the winding affect the available heat transfer. The relationship between the pool boiling curve for helium and the winding configuration is discussed. Strategies for limiting the degradation of heat transfer caused by excessive vapor in the winding are presented. Some closing remarks are made concerning the potential for further improved performance of pool boiling magnets, and areas for additional research of pool boiling/ ventilation in magnet windings are suggested.

The time evolution of a normal zone and propagation of the normal front are calculated for a cryogenically stable superconductor resembling the configuration of a UW Energy Storage Conductor laboratory prototype of 13.5 cm diameter stabilized with 400 RRR A1 carrying 765 kA in a field of 3.5 T. The coupled heat transport and current diffusion equations with temperature dependent physical properties are solved numerically. Both radial adiabatic and superfluid helium cooling cases are treated. A factor of four enhancement in the propagation velocity occurs in the adiabatic case and a transient propagation velocity in the superfluid helium cooling case exists because of the heat generation due to the transient non-uniform current distribuion in the stabilizer.

This report presents measurements on the stability of a He II cooled superconductor. The conductor is helically wound in a phenolic coil support and is cooled by a channel of He II which allows for heat transfer in two dimensions. Stability is measured by recovery of the superconducting state following a transient burst of heat. The data also allows one to measure normal zone propagation velocities and temperature profiles at various times. Currents in the conductor range from 0 to 5 kA with background magnetic field ranging from 0 to 11 tesla. Descriptions of stability boundaries are given in terms of maximum energy bursts from which the conductor recovers as a function of transport current, channel size and background field. Comparisons of the data with an analysis of stability in a one-dimensional channel are shown to be useful. This experimental configuration is intended to model some aspects of a superconducting magnetic energy storage system.

Large superconducting energy storage coils, typically 5000 MWh, require a method of internally dumping their stored energy in an emergency situation. The present scenario envisages: a dump situation is detected; large switch is activated which shorts all conductor turns and shorts the cold structure; liquid helium dump valve is activated; pressurized room temperature helium gas forces liquid helium from cryostat into a reservoir in 5 to 10 seconds; as the turns are uncovered, they go normal but much of the energy is coupled to the other turns still covered with liquid helium and the structure. The system tends to warm up uniformly because of the coupling. Results are presented for the temperature and voltage distribution for the coil with and without the switch. The analysis improves on previous work in accounting for the imperfect coupling between conductor turns and the beneficial effects of having the conductor in electrical contact with the structure, yet electrically insulated from the other turns.

For analyzing the safety and stability of large superconducting magnets, three computer codes TASS, SHORTURN, and SSICC have been developed, applicable to bath-cooled magnets, bath-cooled magnets with shorted turns, and magnets with internally cooled conductors respectively. The TASS code is described, and the use of the three codes is reviewed.

Balanced flow coaxial tube counterflow heat exchangers have application to a number of helium refrigeration systems. Variations in the helium fluid properties in heat exchangers of this type are large enough to cause concern over the accuracy of simplified constant property analyses. The effect of fluid property variations on the performance of such heat exchangers are investigated with a finite difference thermal model. The model uses temperature dependent values of specific heat, thermal conductivity, and viscosity to evaluate heat exchanger performance, but neglects the effects of pressure losses, internal wall thermal resistance and axial conduction. Ratios of analytically predicted to numerically determined heat exchanger effectiveness versus the design N_tu are presented. For values of N_tu of interest in most cryogenic applications (i.e., greater than 3), predicted performance using variable fluid properties is two percent greater than that predicted using mean properties.

Enhanced condensation heat transfer, aided by surface tension, has been measured using vertical tubes with V-shaped grooves on the interior surface. The number of grooves in the same tube diameter were 28,32,36, and 40. The heat transfer rates have been observed to reach high values near the groove number of 40. The results from experiments are in agreement with the theoretical model. For a specified diameter there exists an optimum groove number of V-shaped corrugated system which is found by theoretical analysis.

Experimental investigations of liquid helium heat transfer from a channel configuration are reported. Four model test sections are constructed each consisting of a rectangular cross section channel 127 mm in length and 12.7 mm wide with a narrow gap of dimension 0.5, 1, 2 and 3 mm respectively. The channels are open at both ends to permit natural circulation of the helium coolant. Variables within the experiment include bath temperature and channel orientation. Values for the peak heat flux are compared to available correlations. An empirical fit to the angular dependence of the peak heat flux is suggested.

We report measurements of the time delay to the onset of film boiling at the interface of a superheated solid and liquid helium I in the bath temperature range from 2.2 to 4.2 K. Our temperature dependent results suggest that delay times, t_d, in the millisecond range are related to the applied heat flux, Q_app, by the equation:1$${Q_{app}} = {(a/{t_d})^{1/2}}{\rho _1}{h_{1v}}$$where α(T) is the thermal diffusivity, ρ₁ is the liquid density, and h_1v (T) is the latent heat of vaporization of liquid helium. The origin of this equation is discussed as well as its physical meaning in terms of describing the conditions necessary for the development of a film boiling configuration.

Experimental studies of thermo-syphon flows in radial tubes and loops between the axis and the periphery of a rotating helium cryostat have shown that when heat is supplied at an intermediate radius, the heat is carried radially inwards as A flow and radially outwards as B flow. The results with helium suggest that while the steady state patterns of the A and B flows are complex, the heat is divided approximately equally between the conventional A flow and the reverse B flow.A model of convective heating in the rotating frame is presented and two necessary conditions for reverse convection are identified and discussed. The model predicts reverse convection in liquid nitrogen and this is confirmed by experimental measurement. An array of radial ducts is proposed for the cooling of a superconducting AC generator in order to counter the effects of reverse convection in the helium refrigerant.

The total heat transfer from approximately to 77 K was experimentally studied for a series of different arrangements of multilayer insulation (MLI) on black painted and aluminum taped copper surfaces. The heat flux as a function of the number of MLI layers and of the overall vacuum level was measured. The flux to a painted surface was 24.7 W/m2 with no MLI and 0.64 W/m2 with 30 layers, both at a vacuum of 1.5 × 10−5 torr. The corresponding values to a taped surface were 4.8 W/m2 and 0.52 W/m2. At 1.5 × 10−5 torr, the use of aluminum tape permits one to use approximately one-half as many layers for the same heat flux. The heat flux was measured for six insulation systems from 1.5 x 10−5 torr to ∿ 1 × 10−3 torr. The temperature distribution through the MLI was measured as a function of vacuum level. It was deduced that the apparent thermal conductivity increases with the distance from the cold surface. The effect of cracks in a paint-MLI system was studied by cutting 6-mm wide cracks through a 90-layer blanket. The heat load increased by more than three times the value calculated from the exposed area only.

In the present set of experiments, boiling regimes in saturated He II confined to a vertical rectangular (1 × 10 × 114 mm) channel are studied. These experiments are intended to model a cooling channel in a superconducting magnet. Measurements are made at temperatures of 1.7, 1.9, and 2.1 K at saturation pressure. The various regimes are achieved by changing the steady state heat input. Experimental results include high speed films taken of the helium near the heated surface and within the channel. From the films, the sequence of events for each boiling regime is described. This provides an insight into how one regime evolves to the next as the heat input is raised or lowered. Additional understanding is gained from temperature and pressure measurements of the system.

Flow instability of supercritical helium in curved stainless steel tubing is studied to obtain engineering data applicable to forced cooled superconducting magnets. The tubing is wound in a ten turn coil of 100 mm in radius, the length and inner diameter of the tubing are 3 m and 2.88 mm respectively. The supercritical helium studied is at a temperature from 5 to 11 K, the pressure is 0.5 to 1.2 MPa and the mass flow rate from 0.3 to 0.7 g/s. The results obtained from the experiments are: (1) Wall temperature along the cooling channel under pulsive thermal input is strongly dependent on flow velocity, heat flux, and heating time. (2) Inverse flow has been observed under heat flux high enough for the helium to reach the pseudo-critical line. (3) A linear relation between the energy per oscillation period and the helium pressure has been derived under a DC thermal input to flowing supercritical helium.

A forced HeII circulation loop was tested under heat pulses of varying durations. Integrated enthalpy of each pulse was not preserved and a part of it seemed to go upstream to a saturated Hell bath heat exchanger. Pulse shapes were analyzed and it was demonstrated that the Gorter-Mellink (G-M) relation ceased to hold, at least in the vicinity of the pulse peaks. An average propagation speed, generally proportional to the body flow speed, was detected and analyzed. The results are discussed in conjunction with the theoretical studies of other authors.

A study of forced convection heat transfer in superfluid helium (He II) is initiated to better understand the physical behavior of this process and to compare it with the more familiar He II heat transfer mechanism of internal convection. An experimental assembly is designed to achieve fluid flow by a motor-driven hydraulic pump which utilizes two stainless steel bellows. Each bellows is connected to one end of a copper tube, 3 mm in diameter and 2 m long. The system allows measurements of one dimensional heat and mass transfer where the measured quantities include: temperature profile and pressure drop. The variable quantities are the helium bath temperature, flow velocity and heat input. The helium bath is held at 1.8 K and under saturation pressure. The flow tube is heated at the middle and the flow velocity is varied up to 97 cm/s. The helium pressure is monitored at both ends of the tube and a friction factor is estimated for He II. Temperature measurements are made at seven evenly spaced locations along the tube. The experimental temperature profile is compared with a numerical solution of an analytical model developed for the problem under study.

The heat transfer from thin wires to He II has been investigated at low hydrostatic pressures on ground and at milligravity on board of the NASA KC-135 aircraft. The peak heat flux density at low gravity is found to be less than 10% lower than the lowest laboratory values, the recovery heat flux density, however, about 50%. For the recovery heat flux density a large scatter was observed, indicating a significant drop of this quantity in the milligravity region.

Porous media have become of increasing importance in low temperature applications. In contrast to the usual Newtonian fluid flow near room temperature, the mass throughput characterization appears to be less certain at liquid He4 temperatures, in particular, in the superfluid He II range. In the present work, the Darcy permeability has been evaluated for laminar flow at very small velocities in Newtonian fluids. However, in the non-Newtonian He II superfluid, the Darcy permeability is not so readily obtainable from simple fluid flow experiments. In our experiments, sintered metal porous plugs with a nominal (filtration rating) size of the order 1 µm to 10 µm have been used. The characteristic length diagram for the equivalent Ergun particle diameter, the filtration rating, and the characteristic throughput length has been evaluated as a function of the Darcy permeability. This set of functions is seen to give a comprehensive picture of porous media throughput properties. In non-Newtonian super- fluid He II, the present work relies on the two-fluid model as the frame of reference. It is found that there exists an analog of Darcy’s law for the flow of the normal fluid component of He II. In this analog, the pressure gradient, mass flux, and the shear viscosity are replaced by the thermomechanical pressure gradient, normal fluid mass flux, and the normal fluid shear viscosity.

A 4 K refrigerator and sleeve have been designed to eliminate cryogen boil-off from superconducting Magnetic Resonance Imaging magnets. The refrigerator and sleeve have heat stations that cool radiation shields at nominal temperatures of 77 K and 20 K. A heat exchanger at 4.2 K recondenses helium boil-off. The unit is designed for use in a cryostat which has no liquid nitrogen in the 77 K shield. It has been demonstrated that the refrigerator can be removed and replaced in less than one hour with a loss of only 0.6 L of helium.

Several refrigerators for liquid helium and liquid nitrogen systems have been integrated successfully into IGC manufactured whole body Magnetic Resonance Imaging (MRI) magnet systems. The refrigerators have been tested in systems with magnetic fields of 0.6T to 1.5T. Tests were performed to study the effectiveness of the refrigerators, the magnetic field effects on the refrigerators, the effect of the refrigerators on the field uniformity and magnetic resonance image quality. The interface between the refrigerator and the whole body MRI magnet system cryostat was specifically designed to allow retrofit to the existing IGC magnet systems, while ensuring good heat transfer characteristics and good vibration isolation from the cryostat. The interface between the refrigerator and the cryostat and the refrigerator test results are presented.

In the past five years a considerable amount of clinical test results have become available from Magnetic Resonance Imaging (MRI) and/or Spectroscopy systems based on superconducting magnets with field strengths up to 4 T. This new technique has now proved to be an important diagnostic tool for physicians. Most of the superconducting magnet systems commercially available today are solenoid type magnets cooled by pool boiling of liquid helium surrounded by a liquid nitrogen cooled radiation shield. At DUT we reconsidered the design of the superconducting magnet and the cooling system to get a system that might be more acceptable for use in a clinical environment. We designed a compact 1.5 T double pair magnet system with low temperature cooling system for: a) two radiation shields, b) precooling of the magnet system to 20 K and c) recondensation of the evaporated helium from the coils. The design and first test results of this system will be discussed.

Stochastic cooling, a technique of prime importance for obtaining antiproton beams usable at high-energy colliders, requires wide-band and low-noise signal acquisition and processing. For this purpose, the CERN Antiproton Collector (ACOL), presently under construction, is equipped with six beam pick-up stations operating at cryogenic temperatures. Each station consists of a high-vacuum vessel housing two mobile arrays of pick-up electrodes, cooled to about 100 K by radiation to a surrounding thermal shield, and two preamplifiers kept below 20 K. Refrigeration is provided by a pair of two-stage Gifford-McMahon helium cryogenerators, thermally linked to the preamplifiers and to the thermal shield. The cryogenerators also cool activated charcoal cryopanels, in order to maintain a residual pressure in the vessel below 10−8 mbar. The choice of such cryogenic options is dictated by performance (steady-state and cool-down), simplicity of operation and maintenance, and high reliability.

A hybrid cryostat using a commercial refrigerator and a bath of liquid helium has been built and tested. Its design and characteristics are consistent with operation in a radio telescope. The cryostat cools the superconducting mixer of a sensitive millimetre wave receiver to 2.7 K. The optical axis of the receiver is well defined mechanically, and the temperature of the mixer is stable. The cryostat can operate for one week between liquid helium fills, it is compact and allows easy access to the receiver components.

A Recondensing-type cryostat cooling system, which can supply liquid helium to three machines to test material fatigue at liquid helium temperature, was developed. By the use of a Claude-type refrigerator and distribution of the individually to recondensers installed in the cryostats of the three machines, each of which has a different function, the specimen in each cryostat can be kept at liquid helium temperature independently. During test operation of this system for 500 hours, each cryostat was cooled, warmed and recooled selectively so as not to disturb the stability of the other refrigerating cryostats. As a result, it was proven that the developed system makes it possible to efficiently carry out a diversity of fatigue tests at liquid helium temperature.

In the High Field Laboratory for Superconducting Materials at Tohoku University, three hybrid magnets HM-3,HM-2 and HM-1, in which the maximum fields are 20, 23 and 31 T, have been constructed in the last four years. The cooling systems for the superconducting magnets, SM-3,SM-2 and SM-1, combined with these hybrid magnets are reported in detail. SM-3 and SM-2 which have cryogenic masses of 550 and 3200 kg, were cooled down to about 20 K in 2.5 and 8.5 days, respectively. Each used a cryogenerator with a cooling ability of 50 W at 20 K. SM-1 which has a cryogenic mass of 4700 kg was precooled through an 80 m liquid and gaseous helium transfer line by use of a helium liquefier-refrigerator with a cooling ability of 215 W at 4.5 K. The average heat loss of this transfer line was less than 0.145 W/m.

A liquid helium refrigerated cryopump with a 500 mm inlet flange was designed to operate in the viscous flow regime. A multitude of cryopanels in an accordian arrangement was used to provide a large cryopumping area to accomodate the high heat loads associated with the incident gas flux. Pumping speeds for deuterium gas were measured in the 1.3 × 10−4 to 53 Pa pressure range. In the molecular region the pumping speed was 14,600 L/s. As the pump inlet pressure increased into the transition and viscous flow regimes, the pump speed increased rapidly to a value of approximately 120,000 L/s at a pressure of 13 Pa. Measurements at higher pressures showed a decline in speed as the system becomes heat transfer limited. The observed variations in speed over the range of pressures are related to the transition from molecular to viscous flow, and the heat transfer characteristics of condensing deuterium and boiling helium.

This paper will include an in-depth discussion of the design, fabrication, and operation of the Mirror Fusion Test Facility (MFTF) cryogenic system located at Lawrence Livermore National Laboratory (LLNL). Each subsystem will be discussed to present a basic composite of the entire facility.

The paper describes the build-up of cryogenic supply facilities for JET, the main loads being the neutral beam injector cryopumps. It deals with the LHe and LN₂ supply and distribution system and covers the work through design, manufacture, installation and testing. The cryosupply control employs an autonomous programmable controller operating in conjunction with the JET Central Computer. Prom the Central Computer the operator can initiate actions remotely, using touch panels and mimics, both for the cryosupply system and the cryopump loads. The cryosupply may be operated locally from local control panels. The cryosupply has been fully performance-tested and has exceeded its specification. The first cryopump load has been automatically cooled down and limited integrated tests carried out.

The TORE SUPRA refrigerator design and manufacture have both been governed by several strict design criteria. These include reliable operation over long periods (8,000 hours per year), pulsed thermal loads at 1.75 K and 4.0 K (every 4 minutes), fully automatic control in the various operating modes, low operating costs and acting as a technical demonstration so that larger future designs could be extrapolated from this base. The paper reviews these criteria, and presents the current status.

Performance of the International Fusion Superconducting Magnet Test Facility (IFSMTF) helium refrigerator in Partial Array Tests with three coils is described. The refrigerator was able to cool the coils and facility structure to 4.2 K in 20 days, with maximum temperature differentials of less than 50 K. Boiloff measurements were made for several components; only the lead dewars showed losses substantially higher than expected. Forced-flow cooling tests were also conducted. The coils and facility were warmed to room temperature in 30 days. Several repairs and improvements were carried out. Results of another recent test on the refrigerator alone are reported.

The BNL 24.8 kW refrigeration system is completely installed and major portions of the acceptance tests have been completed. So far, the equipment tested has performed at or above design levels. The room temperature helium compressor station has been completely tested and accepted. The two-stage oil injected screw compressor system exhibited an isothermal efficiency of 57% while delivering a helium flow in excess of 4400 g/s. Data on the performance of the make-up gas cryogenic purifier is also given.The refrigerator turbomachinery, 13 expanders and three cold compressors, has been tested at room temperature for mechanical integrity and control stability. The first cooldown to operating temperature will be attempted in late August, 1985.

A superconducting proton storage ring will be installed in a 6336 m long ring tunnel as part of a 30 GeV electrons, 820 GeV protons colliding beam facility at DESY (Hamburg). The magnets will be cooled by three refrigerators installed in a central cryogenic building. Each unit has a capacity of both 6775 W refrigeration and 20.5 g/s current lead cooling gas rate at 4.4 K, as well as 20 000 W between 40 and 80 K for the cooling of radiation shields. Coolant will be supercritical helium at 2.5 × 105 Pa, T ∼ 4.4 K being continuously recooled by two-phase helium boiling at 1.1 × 105 Pa within the dipoles. A large transfer line going all around the tunnel will supply individual sections. Sufficient redundancy for the most sensitive components ensures continuous operation in case of individual component failures. The whole system will be automated by means of a process control computer system.

Operating experience since the last report, Aug. 1983, is covered. The current mode of supplying liquid helium to the superconducting accelerator ring is explained including a discussion of the interaction of the system with upsets in the ring. The method of controlling helium flow to the dewars, liquid helium pump, helium subcooler, and cold box is very stable and largely automatic. The capacity of the plant is 136 g/s at the current operating point with a peak demand flow of greater than 250 g/s using liquid pumped from dewar storage. The design specifications of the major equipment are tabulated giving the main characteristics of the compressors, cold box, and helium and nitrogen storage. The operating history is analyzed to yield a lifetime efficiency of 97% in 19840 hours of running, and a breakdown of major failures and their causes is given. The major sources of downtime have been contamination of the helium stream by dust, water, and nitrogen. The solutions to these problems are discussed. A new liquid helium pump has been commissioned which has improved the system reliability and performance. This reciprocating pump is described and test data is presented. A third compressor has been commissioned as a backup for the two original compressors. A new control system utilizing a Texas Instruments PM550 programmable controller is used to monitor and control the third compressor. The performance of the PM550 has been excellent, and it is being implemented to control other parts of the system.

The 1500 W helium refrigerator system utilizes two oil-injected screw compressors staged to feed a liquid nitrogen pre-cooled cold box. Refrigeration is provided by two Sulzer TGL-22 magnetic/gas bearing turbines. The refrigerator feeds six magnet test stands via a 10,000 L dewar and subcooler equipped distribution box. The design of the controls has permitted the system to be routinely operated 24 hours a day, seven days a week with only five operators. It operated approximately 90% of the 4 1/2 years prior to a one-year shutdown in 1984, during which the compressor skid was moved. Scheduled maintenance, failures, repairs and holidays are about equal to the 10$ off time. The equipment described was used to test approximately 1200 superconducting magnets for the Fermilab accelerator ring.1 The seven year operating experience is presented as an equipment and technique review. Compressor hours currently exceed 42,000 and turbine hours exceed 39,000 each. Failure rates, causes, preventive maintenance, monitoring practices and equipment, and modifications are examined along with notes on some of the more successful applications of technique and equipment.

A 10-MVA superconducting alternator has been constructed to demonstrate advanced concepts which will permit fault-worthy operation of large central station superconducting generators. The generator is installed in a test facility which includes a 15,000 horsepower gas turbine prime mover and a 50 litre per hour helium liquefier. The facility will permit on-line tests of the generator which will be connected to the local power grid transmitting 13.8 kV power directly into the system. The rotating 4.8 tesla field winding is surrounded by two helium-cooled electromagnetic shields to protect it from heat generated during fault conditions. The design of the rotor cooling system is presented along with a description of rotor instrumentation and data acquisition. Results of rotor cool-down tests are presented.

A saturated vapor helium compressor was designed and tested as a component of a helium-temperature refrigeration cycle. The use of the cold compressor allows reduction of both the precooling heat exchanger area and main compressor size compared to a conventional cycle due to increased pressure of the return gas. The compressor tested was a single-piston reciprocating device which was controlled with programmable hydraulic/pneumatic logic. The compressor was mounted at the cold end of a CTI Model 1400 helium liquefier. An average compression ratio of 2.4 was obtained and an average efficiency of 82% was achieved. In computing compressor efficiency, external heat leaks to the compressor were neglected.

An experimental study of a new type of expander, the Rotary Jet Expander (RJE), has been conducted to advance gas separation engineering and cryogenic engineering. In contrast to reciprocating and turbo-expanders, an RJE is simple in construction and easy to manufacture. According to experimental data obtained, the efficiency of the RJE is high because of low rotational velocity. This, and other characteristic features of an RJE result in a wide range of favorable operating conditions of an RJE. It is also possible to recover the heat energy of the expanding gases. An RJE can operate continuously at low rotational velocities (between 2,000 and 4,000 rpm) of the distributor. High isentropic efficiencies, in excess of 60%, are achieved as long as the dimensions are optimized. The principle of the RJE is described in the paper along with its characteristics, performance values, experimental setup and a prototype design.

Each of the three refrigerators for the HERA proton magnet ring must provide 6.775 kW of refrigeration at 4.3 K plus 20.5 g/s of helium at 2.5 bar and 4.5 K for leads cooling and 20 kW of refrigeration at 40–80 K for shield cooling.The capital cost of large refrigerators is small compared with operating costs. Therefore the refrigeration process was analysed on the basis of exergy. This means the irreversibility of each component is expressed as power input into the plant.The process realised consists of the turbine cycle, divided into two streams with 5 gas bearing turbines all together, and the Joule Thomson cycle. Special attention was paid to the cold end of the plant. The optimization resulted in a new configuration with two turboexpanders running in parallel on different temperature levels.

The capacity of the Bureau of Mines Exell Helium Liquefaction plant near Amarillo, Texas has recently been upgraded by the addition of a reciprocating expander (wet engine) to replace the Joule-Thomson expansion into the two-phase region. The original plant was designed to produce approximately 500 litres per hour of liquid helium from prepurified feed gas using a combination of LN₂ precooling, two oil-bearing expansion turbines and the Joule-Thomson effect for refrigeration. The wet engine addition, combined with higher liquid fraction in the precooling nitrogen feed, led to an increase in liquefaction capacity of 38 percent with no increase in the flow of compressed recycle gas to the liquefier. The modification to the plant is described and its observed performance before and after addition of the wet engine is presented and analyzed.

Comments on the use of two phase helium in a closed circuit tubular cooling system and some results obtained with the TPC superconducting magnet are given. Theoretical arguments and experimental evidence are given against a previously suggested method to determine helium two phase flow regimes. Two methods to reduce pressure in the magnet cooling tubes during quenches are discussed; 1) lowering the density of helium in the magnet cooling tubes and 2) proper location of pressure relief valves. Some techniques used to protect the refrigerator from too much cold return gas are also mentioned.

The cooldown of a bath-cooled superconducting coil was analyzed by both experiment and simulation. The coil, which weighed 155 kg and consisted of 14 double pancake coils, was precooled with cold helium gas. The temperature distribution in the coil, the mechanism of heat transfer and the utilization factor of enthalpy of cold helium gas were experimentally studied. From these experimental results and simulations, effective thermal conductivities of the coil as a compound structure and heat transfer coefficients to helium gas were estimated. As a result, it was clear that thermal conductivities were mainly dependent on the electrical insulator and the helium gas between cables, and that the heat transfer was influenced by the velocity of helium gas adjacent to the coil surface.

The HERA superconducting proton ring contains 422 dipoles, 224 quadrupoles and about 650 correction magnets which are also superconducting. The coils are cooled by single-phase helium in forced flow which is in thermal contact with two-phase helium in counterflow. The total mass to be cooled to liquid helium temperature amounts to about 3.9 × 106 kg. Calculations on steady-state operation and on two methods of cool down and warm up of the magnet system are presented.

Two kinds of applications are considered for magnetic refrigeration, first in the 1.8 K range mainly for He II cooling of superconductors and second for He I refrigeration with precooling near 20 K. At He II temperatures, refrigeration by a Carnot cycle has been investigated and analysed. We give a survey of the results which demonstrate the capability of magnetic refrigeration to attain high thermodynamic efficiency. More interest is now devoted to He I cooling for which we compare magnetic refrigeration to the standard existing solutions. For temperatures higher than 10 K, we discuss the interest and the constraints of cycles with internal heat exchange. Analysis of various configurations is given to explain the choice of a new experiment being built.

A Carnot-cycle magnetic refrigerator has been designed, built, and tested in the temperature range of ~4 K to ~15 K. Gadolinium gallium garnet in the rim of a wheel is the refrigerant. The wheel rim rotates through a gap between two superconducting Helmholtz coils that produce a magnetic field of up to 6 T. Helium gas is used as the heat-transfer fluid in the hot and cold regions of the wheel. The refrigerator performance has been measured in an open-cycle flow system because no suitable low-temperature helium gas pumps were available for closed loop circulation of helium gas. Over one watt of cooling power with a temperature span of several degrees was achieved. At low frequencies the cooling power and temperature changes of the refrigerator match the entropy-temperature data used in the design. Problems associated with friction and gas mixing limit the performance at frequencies above about 0.1 Hz. Separate friction measurements suggest that gas flow control is the dominant problem that needs to be solved before significant improvement in refrigerator operation can be expected. The present measured efficiency is about 20% of Carnot if the drive motor efficiency is ignored. With friction and other losses in the drive motor mechanism, the overall efficiency is ~1% of Carnot.

We have developed a reciprocating magnetic refrigerator for liquefying helium from a temperature in the 15 K region. The working material [gadolinium-gallium-garnet (GGG) single crystal, 30 mm in diameter and 40 mm in length] is placed at the end of the piston. When GGG is placed in a high-intensity magnetic field (4.5 T), its temperature rises to 15 K. The inner surface of the cylinder is cooled by an auxiliary refrigerator; the generated heat is removed through the narrow gap between GGG and cylinder (less than 50 um in the 15 K region) filled with gaseous helium. When the magnetic field in the GGG is eliminated by moving the piston, the GGG temperature falls below 4.2 K and the refrigeration occurs by condensing the helium on the GGG surface.Technical emphasis was placed on the realization of a high heat exchange rate between GGG and gaseous helium. A sufficient heat transfer rate was achieved after several component level experiments. Sources of inefficiency to the refrigeration power has been also discussed.Finally a refrigeration power of 0.95 W at 4.2 K was achieved in 0.38 Hz operation by a reciprocating magnetic refrigerator equiped with two pistons.

We have constructed a special one-shot test apparatus to investigate the properties of a reciprocating magnetic refrigerator for helium liquefaction. The liquefaction efficiency for one-shot demagnetization was only about 40% with an initial temperature of 5 K, and higher initial temperatures gave lower liquefaction efficiencies. Losses in the experiment can be classified into four causes: the thermal resistance of the film condensation heat transfer, friction, the heat flow from the holders to the refrigerant and cooling of gaseous helium above 4.2 K to 4.2 K. The last cause was the most dominant.

The Carnot type magnetic refrigeration in the temperature range from 4.2 K to 15 K has been studied. This paper describes the additional experimental results to improve the liquefaction efficiency. The liquefaction losses in the demagnetization process largely depend on the gap width of the heat pipe-GGG system. At each initial temperature, there was an optimum gap width and magnetic field sweep rate. The maximum liquefaction efficiency was obtained for the gap width of 400 to 600 µm at 4.2 K.

Pulse tube or thermoacoustic refrigerators require only one moving part—an oscillating piston or diaphragm at room temperature. Refrigeration occurs within a tube connected to the pressure wave generator when the thermal relaxation time between gas and tube is comparable to a half period. Three types have been discussed in the literature recently by Gifford, by Mikulin, and by Wheatley. A record low temperature of 60 K was achieved in our work using a single stage pulse tube similar to that of Mikulin. Previously 105 K was the lowest temperature achieved. Because of only one moving part, all three types have the potential for long life, but their efficiency and intrinsic limitations have never been investigated. This paper compares the three types with each other and with common refrigerators such as Joule-Thomson and Stirling refrigerators. An apparatus is described which can measure the intrinsic behavior of the different types from temperatures of about 30 K to 300 K. Overall cycle efficiency as well as sources of loss such as conduction and regenerator ineffectiveness are discussed and the advantages of various phase shifting techniques to increase refrigeration capacity are compared.

This paper describes the development, from a previously proven design approach, of a robust and simple Stirling cycle cooler with long life potential. The need for a closed cycle refrigerator for use in a spacecraft borne infra-red radiometer is explained. The refrigerator is to supply 1 watt of cooling at 80 K for less than 80 watts of input power, be able to survive the launch environment and subsequently run for 26000 hours. Clearance seals achieved with a spring suspension developed from earlier space proven mechanisms have led to the production of a linear split Stirling cycle machine with no apparent life limiting features. A servo control system, in conjunction with moving coil motors and LVDT position sensors, permits running of balanced pairs of mechanisms. The working fluid, helium at a pressure of 1.2 MPa, is contained within titanium bodies having gold O-ring seals. A vacuum bakeout procedure, based upon experience and outgassing trials, reduces residual contaminant release to acceptable levels. A prototype refrigerator has been subjected to a vibration test and has subsequently run for 6000 hours with no detectable change in performance.

The performance of the Oxford Cryocooler is summarised. This cooler has been developed for space use by the Rutherford Appleton Laboratory and the Departments of Atmospheric Physics and Engineering Science at Oxford University. The design goal of 1/2 watt of cooling power at 80 K for 30 W electrical input power has been exceeded by a substantial amount. The power budget for the compressor and the losses in the displacer are discussed. Graphs of the cold end temperature vs. compressor input power and cooling power are presented.

In the field of small size helium refrigerators, need for a maintenance free and vibration free system is rapidly increasing, especially for the application to cryoelectronic devices such as NMR-CT and Josephson computers. To meet this need, we have been developing a Claude cycle helium refrigerator (5W at 4.5K) with two-stage expansion micro-turbines. Other than the micro-turbines, compact size perforated- plate heat exchangers and a single-stage screw type compressor have been developed for the system. This paper describes design and test results of the major components of the system with an emphasis on the micro-turbine performance.

A long-life Joule-Thomson refrigerator which is heat powered, involves no sealing, and has few mechanical parts is desirable for long-term sensor cooling in space. In the gas-sorption J-T refrigerator, cooling is achieved by gas sorption (either adsorption or absorption) processes. Currently, a modular, single-stage refrigerator is being designed and built to be operated at 20 K. The design was analyzed using a dynamic model, which is described here. The model includes the kinetics of the compressors and the heat switches, the heat transfer of the pre-coolers and the heat exchangers, the on/off ratio of the check valves, and the impedance of the J-T valve. The cooling power, the cycle time, and the operating conditions were obtained in terms of the power input, the heat sink temperature, and the J-T impedance.

A small cryogenic expansion turbine has been designed and tested to meet the requirements for reliable and compact helium liquifier-refrigerators. The originality of the turbine is the use of self-acting grooved gas bearings for both thrust and journal bearing. In this case, the construction of the turbine becomes simpler and the rotation is stable at very high speed. Because the radial clearance of the grooved journal bearing is only a few micrometers, it is possible to choose a small gap between turbine wheel and nozzle ring, so that the leakage loss is decreased. Experimental results show that this turbine has advantages of simple structure, reliable operation and good adiabatic efficiency, therefore it has very good prospects for applications.

Development of solid cryogen coolers for spacecraft instrument cooling has been in progress since 1965, with the first orbital operation in 1971. The use of solid cryogens has proven to be very reliable and to assure excellent temperature stability. The six coolers of this type that have been placed in orbit have had lifetimes of 7 to 11 months with cooling to 60 K.More recent developments have extended the operating range with solid hydrogen to 10 K and measured lifetimes to 6–1/2 years. Extension of lifetimes to 10 years or longer can be expected for some requirements (limited cooling loads) by a straightforward extension of existing technology.Work is in progress to improve the efficiency of solid cryogen coolers by reducing parasitic heat loads by improved supports and multilayer insulation system, and by other techniques. This paper summarizes the state of this technology and indicates future trends.

The development of the Active Phase Separator (APS), which was initiated by the rather extreme demands for the GIRL (German Infrared Laboratory) mission, is reviewed. As a final result the APS model V is presented, which is suitable for space qualification and can be considered not only as an universal phase separator but also as a multifunctional component for space cooling systems, allowing a simplification of the valve plan. Besides discussing the scientific results obtained with annular flow channels and presentation of performance tests with the APS, the status of the actual IR telescopes IRAS, (Infra Red Astronomy Sattelite), GIRL, and ISO (Infrared Space Observatory) is summarized.

The porous plug is a proven method for phase separation in superfluid helium dewars operating in zero gravity. The principal operating regimes have been identified in laboratory experiments on porous plugs. In this paper a description of porous plug operation is given, and a new explanation is presented for the observed transition in the m vs. ΔP characteristic. Theoretical results are in qualitative agreement with experimental data.

An experiment to investigate the properties of superfluid helium in a microgravity environment flew on the Shuttle on the Spacelab 2 mission in July and August of 1985. This paper summarizes the flight experiment and describes some preliminary results.The experiment comprised an investigation of long-wavelength third-sound waves in micron-thick films, a study of the motions of superfluid helium under milli-g and micro-g accelerations, and measurements of the fluctuations in temperature associated with the small motions of the bulk helium. An additional objective was to qualify and characterize a reflyable, space-compatible cryostat.

The Infrared Astronomical Satellite (IRAS) which completed the first all sky survey in the infrared demonstrated the tremendous advantage of space-based infrared astronomy. The ability to cool the telescope optics and focal plane to liquid helium temperatures and the absence of atmospheric disturbances which cause “seeing” effects resulted in the discovery of 250,000 IR sources and many interesting phenomena including dust clouds around Vega and the infrared “cirrus” at 100 μm. To realize the true benefit of space infrared astronomy, NASA is now studying the Space Infrared Telescope Facility, a long-life space-based observatory, to follow up on the survey results of IRAS. The choice of orbits is a critical program decision. The objective of this paper is to compare the performance of an all superfluid helium SIRTF system in the two possible orbit inclinations, polar orbit (99°) and the low inclination orbit (28.5°).

The Space Infrared Telescope Facility (SIRTF) is a superfluid helium cooled 85 cm telescope with three infrared instruments at the focal plane. SIRTF will establish in space a long term maintainable infrared observatory for the region of 2–700 μm. The cryogenic system can be designed to last up to six years with 1280 kg of superfluid, and can function in either a 28.5° or 98° inclination orbit by exchanging the sunshade. The lifetime is primarily a function of instrument heat load rather than parasitic heat to the cryogen system.

A preliminary concept of a system for resupplying superfluid helium in space has been developed at NASA Ames Research Center. Such a system would greatly enhance several planned scientific missions. These include the Space Infrared Telescope Facility (SIRTF), Large Deployable Reflector (LDR), Gravity Probe-B (GP-B), Hubble Space Telescope (HST), and Advanced X-ray Astrophysics Facility (AXAF). Resupplying helium in orbit would extend the lifetime for each of these missions without the difficulties, delays, and costs associated with retrieving the system, resupplying on the ground, and relaunching. This is especially true of systems such as the LDR that are assembled in space. A simple thermodynamic model of a transfer system is presented and discussed. The different thermodynamic regimes that can be expected in the transfer line are also discussed. The relations between these regimes are used to evaluate the relative efficiencies of different transfer techniques that could be used with the transfer system. The efficient transfer of liquid helium requires a low heat leak into the transfer line, particularly at point sources such as the coupling. It is also shown that the proper selection of supply tank temperature is critical during helium resupply.

The Cryogenic Fluid Management Facility is a reusable test bed which is designed to be carried within the shuttle cargo bay to investigate the systems and technologies associated with the efficient management of cryogens in space. Cryogenic fluid management consists of the systems and technologies for: 1) liquid storage and supply, including capillary acquisition/expulsion systems which provide single-phase liquid to the user system, 2) both passive and active thermal control systems, and 3) fluid transfer/resupply systems, including transfer lines and receiver tanks. The facility contains a storage and supply tank, a transfer line and a receiver tank, configured to provide low-g verification of fluid and thermal models of cryogenic storage and transfer processes. The facility will provide design data and criteria for future subcritical cryogenic storage and transfer system applications, such as space station life support, attitude control, power and fuel depot supply, resupply tankers, external tank (ET) propellant scavenging, and ground-based and space-based orbit transfer vehicles (OTV).

One approach for increasing the reliability of cryogenic cooling systems using mechanical cryocoolers is to connect several cryocoolers in parallel using thermal switches to isolate failed or standby machines. In this paper, simplified analytical thermal models are used to investigate thermal switch performance requirements and evaluate the effect of thermal switches on system efficiency. Results show that switch characteristics such as switching ratio and thermal resistance determine cryocooler performance requirements. In order to minimize system input power requirements switch characteristics must be carefully matched to the configuration of the cryocooler system.

The development of a thermal switch based on dissimilar materials is described. The switch was developed for use with redundant refrigerators for cooling a gamma-ray detector. It contains no “moving” parts, requires no actuation gas, and is passively switched from off to on or on to off solely by the cooling or warming of the refrigerator cold tip.The switch has been found to be highly reliable while operating down to 80 K (with operation to colder temperature also being posible with design). Four of these switches have been built and tested with completely satisfactory results. Switching ratios of approximately 300 (which include the material resistance of flexible couplings) have been attained for this design, with much higher switching values being possible. This paper describes the switch design and test results.

A cryogenic thermal switch has been designed, assembled, and tested over a range of temperatures. The switch has no moving parts and has a thermal resistance which can vary about a factor of 3000 on demand by the actuation of a heater. The switch resistance can be varied continuously between “on” and “off” values. This device is a gas-gap heat switch filled and evacuated by a heater-activated getter. The switch contains two end pieces constructed from OFHC copper, separated by a thin gap which is filled with helium gas from the getter. The temperature of the getter controls the pressure of the gas in the switch. A thermal model of the device was set up to predict its performance for various temperatures and gas pressures. Tests down to liquid-helium temperatures verified the predicted resistance ratios between “off” and “on”.

Passive orbital disconnect struts can potentially reduce the support conductance a factor of 10 over state-of-the-art tension band nondisconnect supports and cut helium dewar weights in half for the same lifetime. A series of thermal and structural tests were performed to verify that this performance improvement is real. Structural tests on a PODS-III support system (consisting of six struts) is reported here. The results show the predicted performance improvements can be achieved and the PODS-III supports are ready for flight applications. For large tankage systems requiring higher side load capability, a PODS-IV version is currently being developed.

Preliminary designs for the NASA/Stanford GP-B Relativity Gryoscope Experiment indicate the desirability of having a capability for assembling and disassembling connections at 4 K which are leak tight to superfluid helium. A series of experiments were performed to prove that this is feasible. Three flanges were tested: a 6.4 cm diameter ConFlat flange, a 7.6 cm diameter indium sealed flange and a 25.4 cm diameter indium sealed flange. The 6.9 cm diameter and 7.6 cm diameter flanges were bolted together while submerged in liquid helium and leak checked while the helium bath temperature was lowered below the λ point. The 25.4 cm diameter flange was bolted together and leak checked at 76 K. All three flanges were leak tight, after being assembled, to a level of less than 4 × 10−7 standard cm3/s.

The number of cryogenic storage system applications is growing with ground, sea and space uses. Fluid stratification is an area of concern for both subcritical and supercritical storage. Stratification is an nonequilibrium event which occurs when heat is locally added to the fluid. This paper presents a simple three-node approach to a test analyzing tank stratification and compares that models1 results to ground data. Ground data were collected as part of the Beech Aircraft liquid methane vehicle conversion program.

Recent interests in hydrogen energy systems have prompted research activities in the safe and effective storage and handling of liquid hydrogen. Small cryogenic tanks will be required for use in future on board vehicular and remote site hydrogen based systems. In order to facilitate design and safety analyses, the Institute for Hydrogen Systems has developed a numerical model which simulates the operations of a small size cryogenic tank. The simulation is based on a simplified physical model in which the mechanism of heat transfer is represented as a one dimensional process and the fluid mixture is approximated as a saturated homogeneous mixture. It is shown that the numerical results, when compared with experimental data, give rise to a prediction of pressure rise in a closed vessel which is 1.3 times lower then the actual pressure rise in a nearly empty tank. The calculated pressure rise is 1.6 times lower for a nearly full tank. This suggests that the difference between the predicted and experimental results is due mainly to the development of temperature gradients in the liquid.

If the vacuum vessel that insulates a cryogenic dewar for a spaceborne experiment prior to launch is damaged, air will leak into the vacuum insulation space. As the sudden heat load causes the pressure to rise in the dewar, a safety disk in the emergency vent line will burst at the design pressure differential to allow vaporized cryogenic fluid to escape. The emergency vent line should be sized such that sufficient gaseous cryogen will be vented to keep the pressure inside the dewar below the design limit. On the other hand, the line should not be so large as to impose an unnecessary heat load on the dewar filled with cryogenic fluid. A vent line computer program was generated to compute the maximum flow rate allowed for a proposed vent line system. Parametric studies have been carried out for different burst disk pressure differentials, liquid cryogen ullage, and vent line sizes. These programs are useful for sizing an emergency venting system for a cryogenic dewar.

This paper reports on the progress of an in-house research program to deve-lop the technology required to use composite honeycomb construction for the outer shells of cryogenic dewars for aerospace use. Fabrication and testing phases are described. Results indicate that high quality, consistently performing compo¬nents can be constructed using this type of construction. Fabricated honeycomb outer shell hemispheres weighed nearly half as much as the current state of the art all-aluminum version. An aluminum foil permeation barrier has been devel¬oped that, on a material sample basis, demonstrates a helium leak tight inter¬face. External pressure collapse tests validated the structural capability and con¬sistency required for use in space.

This paper describes the development of a second-generation centrifugal circulator for cryogenic fluids. The circulator is designed to operate over a wide range of flow rate and pressure rise and can be used for the pumping of liquid and compression of vapor at temperatures down to liquid helium (4 K). The machine incorporates self-acting gas journal bearings, a permanent magnet axial thrust bearing, and a variable speed induction motor drive to provide for reliable, maintenance-free operation. The paper will describe design details of the pump. Calculated performance characteristics for a liquid helium pumping application will be presented along with a general discussion regarding limitations of the present system.

Submersible electric motor driven fans of three sizes have been designed, built and operated at 21 K at the Stanford Linear Accelerator Center. The largest is a 100-mm diameter, 2 stage vaneaxial fan with a nominal capacity of 6 L/s at 2 m head. It is driven by a 4 pole, 3 phase induction motor that runs at 1750 rpm. The next smaller one is an 85-mm diameter centrifugal pump. It pumps 3 L/s at ahead of 5 m. The third is a 75-mm single stage vaneaxial fan with a nominal capacity is 3 L/s at a head of 2 m. The 85-mm pump and the 75-mm fan are driven by 2 pole, 3 phase induction motors running at 3550 rpm. The motors were modified to operate submerged in the cryogenic fluid. The pumps have been operated in liquid hydrogen, liquid deuterium, and pressurized helium gas at 21 K. They can also operate with denser fluids such as liquid nitrogen, but rotational speed, capacity, and head will be reduced. They have been operated while submerged in liquid helium.

In a refrigerated seal the fluid to be sealed flows between a rotating shaft and a refrigerated housing. The fluid can be frozen to the housing during the transient phase. Under steady-state conditions the refrigerated seal proved to be a dynamic low-leakage seal. The concept was extended to pressure differences of 6.9 MPa (1000 psi).

Special transfer lines were required to transport supercritical hydrogen gas at 20 K and 1.5 MPa and liquid methane at 800 K and 0.4 MPa over a distance of about 18 m, to moderator vessels sited within a concrete shielded area.The moderator vessels (1 L and 1/2 L volume respectively) are built into shielding material surrounding the uranium target which produces high energy pulsed neutrons. Energies up to 400 W can be deposited in the cryogenic fluid, which the control system must be capable of handling within less than one minute. The flow rates were designed to be 500 cm3/sec for hydrogen and 220 cms3/sec for methane. In order to maintain biological shielding integrity the lines had to be as compact in cross section as possible, and replacable using remote handling techniques. As commonly used plastic materials could not be used for spacers etc, on account of their poor radiation resistance, new techniques had to be developed to meet the long term requirements of the lines.

A liquid helium transfer line system was designed and constructed for a prototype 3-cell superconducting cavity (SCC) which was Installed in the TRISTAN Accumulator Ring (AR) for electron beam acceleration test. The liquid helium transfer system, about 20 m long, consisting of three helium lines (supply, return and bypass) and two liquid nitrogen lines (supply and return), connects the helium refrigerator or liquid helium dewar at ground level and the SCC cryostat in the underground AR tunnel through a narrow service channel. The operational modes (cooldown, liquefier and refrigerator cooling mode) can be selected by valves in the transfer line system. The system worked very well and the SCC was cooled and tested successfully.

Hydrogen fueled internal combustion engines show promise in reducing polluting exhaust emissions. Internal cryogenic mixture formation especially enables exhaust emissions below the limits of existing regulations. Injection pressure requirements for internal mixture formation still require high pressure liquid hydrogen fuel pumps. LH₂ can be handled and stored onboard a vehicle similar to other fuels such as liquefied petroleum gases and liquefied natural gas. LH₂ fuel tanks suitable for automotive application are now state-of-the-art. Tank refilling can be made safely by automatically operated refilling stations. Two research activities have been started between DFVLR and two major automobile manufacturers in Germany. Different types of mixture formation and engine conversion for LH₂ fueled automotive vehicles are the subject of this R&D. By adjoining the heat sinking capability of LH₂ by cryogenic mixture formation to the the internal combustion engine, performance can be improved considerably. This seems to be the most efficient measure in the development of optimized hydrogen fueled internal combustion engines.

A passenger car with 2-L four-stroke turbo-diesel engine was converted into a LH₂-fueled vehicle. The hydrogen fuel system consisted of a LH₂tank and a high pressure LH₂pump. The engine used 8 MPa H₂ late direct cylinder injection and hot surface ignition. This system demonstrated the advantages of light weight and high performance as an automobile hydrogen engine. It was superior to a system using a metal hydride tank and premixture type engine.

This paper summarizes the results of two projects and the work of the Canadian Hydrogen Safety Committee (HSC). The first project covered the safety of hydrogen as a ground transportation fuel. The preferred form of hydrogen was as cryogenic liquid although other forms were also considered. Designs for bulk storage, distribution, retail storage, refuelling and in-vehicle use were developed and the hazards identified by “Hazard & Operability Studies”. The second project was concerned with finding “niches” where the unique qualities of hydrogen as an energy- carrier could be exploited in the near future. The study covered a wide range of “applications-driven” uses including utilities, road and rail transport, aircraft, mine vehicles and submarines as well as “technology-driven” options such as the exploitation of low-temperature effects.

In the last 20 years cryogenic air separation plant sizes have increased from 150 tons per day to 2800 tons per day. Progressively reversing heat exchangers have replaced regenerators. However, with this arrangement, the quantity of pure product output is limited to about 50% of the air input. With the appearance on the market of molecular sieve, another arrangement was developed, which allows one to produce a combined pure product flow equivalent to 85% of the air input. Recently, there has been a strong tendency for the reversing exchangers to be superseded by this arrangement.Due to the ever increasing cost of energy, optimization studies are today pushed much further than they used to be; as a consequence there have been major changes in the size of equipment, improvements in the machinery, and simultaneous developments in instrumentation.

Over the past twenty years separation processes based upon pressure swing adsorption have replaced cryogenic processes in a number of selected applications such as air separation for production of moderate quantities of nitrogen and oxygen and recovery of hydrogen from refinery and chemical plant gases. Key events contributing to the emergence of PSA as an important process option have been the development of synthetic zeolite molecular sieves by Union Carbide Corporation in the USA and of carbon molecular sieves by Bergbau-Forschung in Germany.Today PSA processes enjoy significant commercial use producing oxygen from 0. 1 Nm3/h for medical application to 1500 Nm3/h for steel mill use, for making nitrogen up to 1000 Nm3/h for inerting and in purifying hydrogen streams of up to 100,000 Nm3/h for refinery use.In this paper some of the principles of adsorptive separations are reviewed. The history of the technology is traced briefly with emphasis on key material, process and application events. The major commercial processes in the application of adsorption to bulk separation of air and hydrogen purification are reviewed in more detail with comparisons made to cryogenic alternatives in terms of specific characteristics, advantages and disadvantages where appropriate. Information on performance, reliability and comparative economics are discussed where available.

The Fermilab Nitrogen Reliquefier (NRL) has been designed to provide the liquid nitrogen requirements of the Energy Saver; Central Helium Liquefier, helium transfer line, the superconducting accelerator magnets and other components. The evaporated gas is collected in a system of pipes, liquefied, make-up liquid added, and recirculated. The design is physically compact and cost competitive with all alternate means of supplying liquid nitrogen refrigeration, including air/separation liquefaction. The T-S diagram, process conditions, and commissioning experience to date are described.

Since 1964 the production capacity of LNG has grown at the rate of 20 percent per year worldwide. The large number of projects currently under consideration forecasts significant growth in the future. LNG production from a unit train has more than doubled since 1972 resulting in a reduction in the installed cost of a unit train of approximately 30 percent. The efficiency of the liquefaction process is now 35% greater than in 1972. The propane precooled-mixed refrigerant liquefaction process has been used in essentially all of the LNG projects since 1972. LNG process and equipment components will be improved incrementally in all areas in the future. Extensive application of modular construction techniques will reduce the time and cost of construction in remote areas of the world.

Operators of all LNG plants must meet certain requirements of the Federal Safety Standards1 (Standards). Although plants existing at the time these Standards were adopted need not be updated to comply with siting, design, installation, or construction requirements, their fire protection and security systems must be updated to comply with these Standards, They must also be operated and maintained in accordance with the Standards, including preparation of procedures, and training of personnel. Some operators use a more advanced technique - a system hazard analysis, which goes beyond the requirements of these Standards - to enhance the reliability and safety of their plants. In this paper we summarizes some techniques for evaluating plants, and the measures taken by some operators to update their plants.

In order to simulate geysering phenomena in LNG pipelines, a system composed of a vertical pipe and a long horizontal one was filled with Freon-113 as the test medium. When the system was heated by electrical heaters along the tube walls, geysering was observed under some conditions. Parameters, including feed system, pipe wall heat fluxes, and liquid velocity, were varied. The effects of these factors on geysering were studied. It was found that the geysering was characterized by the storage and release of thermal energy in the liquid column and in the long horizontal line. The conditions under which geysering occurred were obtained experimentally. The region where geysering occurred was shown as a map of heat flux and liquid velocity.

A 140,000 m3 LNG inground storage tank is now being constructed at the Tokyo Gas Sodegaura LNG receiving terminal. In order to reduce costs, we made improvements over the 130,000 m3 tanks constructed previously. Improvements in the slurry wall included utilizing it as the major component of the side wall by using high strength underwater concrete with a high performance water reducing agent, modifying the circumferential joint and by integrating the slurry wall with the side wall by binding rebars. The slurry wall itself can withstand earthpressure and groundwater pressure during construction and the composite side wall can withstand seismic and thermal loads after completion. We also adopted a drainage bottom type structure based on actual seepage-water flow rate of the 100-m deep slurry wall developed and constructed for 130,000 m3 tanks. In addition to these, this tank is designed to receive LNG from different source by adopting countemeasures such as top/bottom feed, densitometer, etc., against roll-over phenomenon.

In recent years close attention has been given to increasing the integrity of LNG storage tanks. The M.W. Kellogg Company is a participant in four major LNG projects that incorporate enhanced integrity LNG storage tanks. Enhanced integrity tanks provide more secure containment than the conventional, above-ground, double-wall steel tank. These projects, the Belgium and Hsing-Ta, Taiwan receiving terminals and the Das Island, and Burrup Penninsula, Australia liquefaction plants each utilize a different approach to the design of enhanced integrity storage.This paper discusses the pertinent features of the primary and secondary containment systems for these designs and compares them with a conventional above-ground, double wall system. Relative costs and design/construction schedules compared to the conventional, above-ground, double-wall steel tank are also reviewed.

Foundation heating systems for LNG storage tanks resting on grade are important in preventing the possibility of frost heave beneath the tank foundation. Here the thermal behavior of such systems is considered by determining how critical parameters influence the design of a heating system. These parameters are: heater spacing, heater cable characteristics, insulation thickness, product temperature, and thermal conductivities of the heater embedment material and foundation soil. The transient behavior of an on-off type of heating control system is also analyzed.Finite element analysis results indicate that the heat transfer required to maintain the foundation at a set temperature, is directly proportional to the heating cable temperature. This relationship is dependent primarily upon the thermal conductivity of the heater embedment material. Knowing this, one can determine in a simple manner the operating temperature of a heating cable as well as the heat output for a selfregulating type cable.To ensure proper heat distribution within the foundation certain areas of the tank foundation require special consideration. These are areas of discontinuity within the bottom insulation system as found near tank bottom penetrations and the concrete ringwall.The selection of a suitable electric heating cable must be considered to optimize the design of such systems. Some of the advantages and disadvantages of four types of electric heating cables are listed. The four types are: (1) constant watt, (2) mineral insulated, (3) self regulating, and (4) single conductor electrical wire.

In storage facilities accepting LNG, it is important to predict the rate of the abnormal increase of boiloff gas at the time of the occurrence of rollover. This report describes the engineering simulation developed by relating known heat transfer phenomena. The fundamentals of a model are based on the stratification and natural convection of the fluid in a vessel, and also on the mixing process. It is assumed that the transfer of heat and mass depending on natural circulation and local disturbance is carried out. The model is, so to speak, a “natural convection model”.

A new fundamental equation for correlation of thermodynamic property data for fluids is presented. The fundamental equation (equation of state) is explicit in Helmholtz energy and is readily adapted to system analysis applications. All thermodynamic properties are derived by differentiation of the fundamental equation. A comprehensive function containing up to 100 terms provides the basis for the correlation. The fundamental equation for a specific fluid is a subset of this comprehensive function. The individual terms of the comprehensive function may be easily changed by varying exponents of the functions of the independent variables. Functions for the calculation of derivative properties are given, and the incorporation of calorimetric information via ideal gas heat capacity equations is discussed. Applications to fluids of cryogenic interest include oxygen, nitrogen, argon, ethylene and neon. Coefficients for calculation of thermodynamic properties of these fluids taken from formulations published elsewhere are given.

A three-parameter pseudo-cubic equation of state is proposed. The pseudo-cubic equation of state is a sixth power polynomial equation in volume, which may be solved in a manner similar to conventional cubic equations. The proposed equation has a maximum of three positive roots. The proposed equation has successfully described the critical isotherms of argon, carbon dioxide, and water. The three adjustable equation-of-state parameters have been treated both as a function of temperature and as constants. Performance comparisons with other three parameter equations of state are given. The results of the comparisons indicate the proposed equation or state to be superior in a number of ways, especially in the prediction of liquid densities. The proposed equation of state has also been successfully applied to the correlation of high pressure vapor-liquid equilibria of cryogenic fluid mixtures.

The temperature dependence of the cohesion parameter “a” (=Ω_aRT_C2 /P_c) of the van der Waals equation of state has been established for predicting vapor-liquid equilibrium (VLE) values for hydrogen-containing systems at cryogenic conditions. Instead of using one temperature function for Ω_a of hydrogen and another for other components as suggested by Graboski and Daubert, a single temperature function for Ω_a was developed for components of cryogenic interests including hydrogen. The optimal Ω_a values of hydrogen above its critical temperature were determined from VLE values of binary hydrogen - n-alkane systems without using the conventional binary adjustable parameter k_ij, and correlated. The success of the proposed correlation was demonstrated by predicting the VLE values for hydrogen-containing systems, including four binaries with the second component other than a n-alkane, and two ternaries. The van der Waals equation used in the calculation is the simplest cubic equation of state. For the systems considered, the agreement obtained between the calculated and the experimental K values from the proposed correlation is about the same or even better than those obtained from the method of Graboski and Daubert.

The published experimental data on the thermodynamic properties of neon have been used as the basis for a new thermodynamic property formulation for neon. The new correlation uses a fundamental equation (equation of state) explicit in Helmholtz energy, which provides for the calculation of derived thermodynamic properties by differentiation. The fundamental equation for neon is a subset of a larger comprehensive function which has also been used in developing thermodynamic property formulations for other fluids of cryogenic interest including oxygen, nitrogen, argon, and ethylene. In addition, new equations for the vapor pressure, saturated liquid density, and saturated vapor density are presented. The formulation presented here may be used to calculate pressure, density, temperature, enthalpy, entropy, internal energy, isochoric and isobaric heat capacities, and velocity of sound for neon. Summary comparisons of properties calculated with the new formulation for neon with selected experimental data are included to verify the accuracy of the fundamental equation for calculation of thermodynamic properties.

Measurements of the orthobaric liquid densities and dielectric constants of carbon dioxide have been obtained at temperatures between 220 and 300 K. Densities were determined with a magnetic suspension densimeter, while a concentric cylinder capacitor was used for. measurements of dielectric constant. The experimental densities and dielectric constants have been used to compute values for the Clausius-Mossotti function. Comparisons with the experimental results of other investigators are presented.

An effort is under way to produce deuterium triple point sealed cells for a 18.7 K temperature reference point. Stability with time of impurity content in cells, namely HD, was found to be good from measurements made at IMGC since 1978, but nearly all commercially produced “chemical pure” D₂ contains approximately 0.5% HD. To reduce the HD content in the D₂ cell, two steps have been taken: (1) The stainless steel cell was chemically etched and then vacuum baked to remove Fe from the surface and to diffuse protium out of the cell. (2) The cell was flushed and filled to 70 bar with D₂ directly obtained from a thermal diffusion column with HD <70 ppm. Triple point measurements with the cell show an initial drift due to para-ortho conversion of 0.4 mK per hour decreasing to 0.07 mK per hour after 130 hours of conversion. The initial triple point temperature on NBS-IPTS-68 was found to be 18.732 ± 0.001 K; this agrees with the value found in the earlier IMGC cells, filled with commercial D₂, when the latter are corrected for a 0.4 ± 0.1 HD content, but the conversion in this cell was 10 times faster than in 304 stainless steel IMGC cells. Further measurements have been made using the clean n-D₂ in an unetched 304 stainless steel cell which has been flushed with D₂O vapor.

The low level signals from cryogenic sensors and transducers are usually carried to the electronic signal conditioning and data handling systems at ambient temperatures by long electrical leads running from the cryogenic environment to ambient. There are many applications, outside those using superconducting devices, in which there are advantages to be gained by placing part or all of the electronic system in the cryogenic environment adjacent to the measuring point. This paper discusses the requirements for an ideal cold electronic instrumentation system and then reviews the present state of the art in relation to off-the-shelf electronic components, devices and integrated circuits, and the published literature. The integration of sensors/transducers with cold electronics is discussed and areas for development are outlined.

For the measurement of temperatures between 1 K and 300 K, there are available a wide variety of sensors and several dc, ac, and pulse techniques. The choice of measurement system depends on a number of factors, including the required precision, cost, speed of response, sensor size and availability, and the effect of high magnetic fields. When the most significant of those factors have been identified and quantified, it should then be possible to select the thermometer(s) that best meet those requirements. To aid in the selection procedures, some pertinent characteristics of a number of useful thermometers are presented. A brief discussion of high magnetic field sensors useful at low temperatures is included.

Recent experiments on exotic heavy Fermion superconductors, superconducting organic conductors, metal-to-insulator transitions in solids, and on the two-dimensional electron gas in semiconductor devices (quantum Hall effect) have needed a combination of low temperatures (T < 0.3 K) and very high magnetic fields (B ≤ 23 T). The low temperatures are achieved with 3He–4He dilution refrigeration, and the high fields are obtained with resistive water-cooled Bitter solenoids powered by large DC generators. In the range 23 ≤ B ≤ 30 T a hybrid magnet arrangement consisting of a superconducting outer solenoid and Bitter inner solenoid is used. Detailed studies of the temperature dependence of the properties of the systems under investigation have made necessary increasing accuracy in temperature determination to distinguish between sometimes subtle differences in theoretical predictions. Two principal difficulties which arise in these investigations are: heating due to changes and fluctuations in the magnet current, and a strong magnetic field dependence of most temperature sensors below 1.0 K. These two factors combine to change both the temperature of the system under investigation and the calibration of thermometer. Present methods of temperature determination (vapor pressure, paramagnetic salts, calibrated resistors, capacitive transfer standards) will be discussed, and plans for primary thermodynamic thermometers with known magnetic field dependence will also be presented.

The measurement of transient temperatures in cryogenic fluid flow requires a highly sensitive, intrinsically fast sensor that is in good thermal contact with the fluid but in poor thermal contact with the solid walls confining the fluid. A resistance thermometer made from a 1 µm thick silicon layer on a 125 µm thick sapphire substrate has a calculated intrinsic response time of about 10 ns at 4 K, and its sensitivity is comparable to germanium or carbon thermometers in the range of 1 – 80 K. This paper describes a novel construction method to mount the small silicon-on-sapphire thermometer in an oscillating fluid flow. The large surface area of the thermometer provides good thermal contact with the fluid, while the suspension ensures poor thermal contact with the holder, maintains its fast response time, and withstands high velocities and frequencies of fluid oscillation. A self-heating response time of 300 ns was measured at 4 K in liquid and gaseous helium. Repeatability of the thermometer is ± 10 mK at 4 K. Examples of the performance of this thermometer for helium gas oscillations in the frequency range of 1 Hz to 12 kHz are given.

A review of diode thermometry is given, pointing out its advantages and limitations. Research and development efforts towards improving the diode temperature sensor are outlined and preliminary data are presented on a recently introduced diode temperature sensor made of GaAlAs. Important aspects to consider in the calibration of temperature sensors and also the limitations in using liquid cryogens as calibration check points are described.

Thin film temperature sensors for use in the liquid helium temperature range have been fabricated from an amorphous semiconductor-metal alloy of germanium and copper. The amorphous material (0.5 – 1.0 µm thick) is sandwiched between two thin film metal electrodes. The active area of the sensor is between 0.1 – 1.0 mm square. The resistive amorphous material is protected from exposure to light and the atmosphere because it is sandwiched between the electrodes. The sensors can be fabricated on any smooth, non-conducting surface which can be placed in the vacuum chamber of an evaporation system and heated to about 80°C. Thin films of silicon oxide insulation have been used to allow fabrication of sensors on a sheet of stainless steel foil. Non-ohmic conduction begins for electric fields exceeding about 10,000 V/m. A transition between conduction mechanisms at low temperatures or high electric fields produces a region in which the resistance is almost inversely proportional to temperature when measured with a constant excitation current. The magnetoresistance, measured from 0–0.6 tesla at 4.36 K, is negative and obeys a power law of the form △R/R = a Bn where exponent n is 0.72. At 0.6 tesla, △R/R is −1.3%.

The design and construction of a noise thermometer built around two commercially available SQUID systems is presented. The thermometer is based upon the direct-coupled system of Webb, Gifford and Wheatley but uses a novel digital squarer/integrator to analyse the noise voltage. Preliminary results demonstrate the performance of the thermometer.

We performed a experiment to detect a temperature rise at cryogenic temperature using a dual-core optical fiber. This fiber has two single-mode optical cores in one fiber. We demonstrated that a temperature rise of 4 K was detectable at 4.2 K. The sensitivity of this method can be improved using a longer fiber. This method may be applicable as a quench detector for superconducting magnets. A quench detector using this optical method is immune from electromagnetic noise, free from troubles caused by break-down of electrical insulator, and has many advantages over a conventional quench detector measuring voltages of a magnet.

Precise pressure measurement is an important requirement frequently encountered in cryogenic experimentation. Applications range from vapor pressure and gas thermometry to gas sample density measurements. Since a capacitance-type pressure sensing element can be constructed of non-magnetic, UHV-compatible materials to cover a wide range of pressure, be bakeable to 400°C, and be reliable and easy-to-operate, it is an ideal candidate for this task. Design and operational characteristics of this family of instruments with reference to low temperature applications will be reviewed.

The heatmeter or more precisely the heat flow meter is a device based on thermal conductivity measuring tecniques. It consists of a thermal conducting body between two thermometers, with provisions for thermal connections and calibration. The paper describes the design, construction details, calibration and performance of one such heatmeter for use with heat reservoirs at either 4.2 K or 78 K. Near 4.2 K it has a sensitivity of 10 µW and requires less than a minute to reach steady state. Near 78 K its sensitivity is 10 mW and requires 25 minutes to reach steady state.

Heat leak measurements of superconducting magnet suspension systems, and multi-layer insulation (MLI) systems are important for the optimum design of magnet cryostats. For this purpose, a versatile cryogenic test facility was developed having a functional end in which test components of differing geometrical configurations could be installed and evaluated.This paper details the test facility design and operating parameters. Experimental results of heat leak measurements to 4.5 K obtained on a post type support system having heat intercepts at 10 K and 80 K are presented. Included are measurements obtained while operating the 10 K intercept at temperatures above 10 K, i.e., in the 10–40 K range. Also reported is a description of the test facility conversion for a heat load study of several MLI systems with variations of MLI installation technique. The results of the first MLI system tested are presented.

This report describes a system to measure the quality factor X in a two-phase helium flow. The two-phase helium flows through a concentric cylindrical capacitor governing the frequency of an oscillator placed at low temperature. The difference between the dielectric constants of liquid and gaseous helium results in a capacitance change, when the ratio of liquid to gas varies in the stream. The capacitance probe is optimized for minimum pressure drop and flow perturbation. Two different modes (analog and digital) are shown to obtain a reading of the quality factor with a precision of about 1% of full scale. The oscillator has excellent stability and reproducibility with respect to temperature variations. The measured quality factors are compared with the results obtained by separate gas flow and liquid helium measurements after phase separation.

An accurate measurement of direct currents in closed superconducting circuits is often difficult to perform, especially if the circuit is exposed to time dependent and/or high stationary magnetic fields. In this paper a new type of non-contacting direct current meter that can measure large currents (in this case up to 25 kA) with an accuracy of 0.1% of full scale is presented. The device consists of a superconducting toroidal transformer that encloses the unknown current flowing through the aperture, a shielded field sensing coil which is connected to the transformer but positioned in a weak magnetic field, and a control unit that presents a reading of the measured current.

A multiplexed temperature data acquisition system with an overall precision of ±25 PPM has been designed using state-of-the-art electronics to accurately read temperature between 2.4 K and 600 K from pre-calibrated transducers such as germanium, silicon diode, thermistor or platinum temperature sensors.

The possibility of using second sound holography to make images of objects inside superfluid liquids has been discussed by various authors for many years. An attractive technique is that of using deformations of the free surface of superfluid helium to measure the second sound amplitude. In earlier experiments to test the feasibility of this, we showed that pictures of the surface deformation due to standing waves could be obtained. We have now been able to show that the boundary conditions imposed by the need for irrotational flow of the superfluid limits the steepness of the allowed steady surface waves. Non-linearities in the relation connecting surface slope and second sound intensity become important close to this limiting steepness. The possibility of second sound holography remains if sufficient stability and observational sensitivity can be combined.

New types of cryoprobes have been developed for the cryodestruction of varicose veins and internal lesions such as bronchial, gastric, colonic tumors or oesophageal tumors or varices. These probes can be introduced inside the operative channels of commercially available endoscopes dues to their small diameter (2 or 3 mm) and flexibility. The operating principle is J.-T.expansion of nitrous oxyde inside the tip of the probe where a heat exchanger and self controled flow to avoid cooling down of the feeder. Furthermore the probes are designed for monitoring the size of the frozen area by means of the bioelectrical low frequency impedance method which is essential for correct use and standardisation of the procedure. The probes are described and examples are given of the field of application.

Recent advance in therapeutic gastrointestinal endoscopy and cryosurgical techniques have caused an interest in creating a cryo-instrument for endoscopic use. The development of a prototype instrument for use in both upper gastrointestinal endoscopy and colonoscopy is described. This instrument is based on the expansion of high pressure carbon dioxide in a series of flexible tubes designed to fit through a 2.6 mm endoscopic accessory channel. This self-enclosed system has been used clinically for cryo-extraction of cauterized colonic polyps. Its potential use for treatment of gastrointestinal malignancies is discussed.