The concept of a nuclear rocket engine is simple and consists of using a nuclear reactor instead of a combustion chamber for heating gas producing the thrust during its escape from a supersonic nozzle. The efficiency of a rocket engine is determined by the ejection rate of gas from the nozzle, which is inversely proportional to the square root of the molecular weight of the gas. Hydrogen has the smallest weight (2 a. m.u.). The ejection rate of hydrogen heated up to 3,000 K is more than twice that of the best chemical fuels for LREs. This is an advantage of the NRE, which can heat pure hydrogen (the mean molecular weight of combustion products in LREs always exceeds 10 a. m.u.). Instead of the ejection rate, the engine efficiency is often characterized by the specific thrust, equal to the ratio of the momentum imparted to the rocket engine to the mass flow rate of the working substance. (The specific thrust is also defined as the ratio of the thrust to the weight flow rate of the fuel and is measured in this case in seconds.) According to calculations, the mass delivered from a circum-terrestrial orbit to the geostationary orbit by an interorbital spacecraft equipped with an NRE having a specific thrust in the range from 850 to 4,400 s is three times larger than the mass delivered by a spacecraft with an LRE.

Aside from these important advantages, the NRE has substantial disadvantages. The main disadvantages are as follows. First, the NRE reactor is considerably heavier than the LRE combustion chamber. Second, the reactor is a high-power radiation source requiring a radiation shield. This makes the engine even heavier and consid­erably complicates its operation. The requirement of the stability in hydrogen at high temperatures and neutron-physics restrictions strongly reduce the choice of materi­als that can be used for manufacturing HREs and high-temperature elements of the HRA. Because the NRE is an air-borne reactor, it should be light enough. Therefore, this reactor should provide a very high energy density, exceeding the energy density of ground energy reactors by a few orders of magnitude [1, 2].

A number of important principles for using brittle carbide materials in HRAs were formulated during the construction of the HRA. The HRA should be made of functionally autonomous units and remain operable even if some of its units are damaged. Each unit is made as a technologically independent aggregate that does

A. Lanin, Nuclear Rocket Engine Reactor, Springer Series in Materials Science 170, 9

DOI: 10.1007/978-3-642-32430-7_2, © Springer-Verlag Berlin Heidelberg 2013

Fig. 2.1 Principal HRA scheme: 1 nozzle unit; 2 bearing grid; 3 beryllium-steel bilayer hous­ing; 4 heat-insulating packet; 5 heating units; 6 input grid; 7 end deflector; 8 thermal-expansion compensation unit; 9 throttle

Fig. 2.2 Longitudinal and transverse sections of the IVG-1 reactor

not require complicated connections with adjacent units during its mounting into the assembly.

Such a functional, technological assembly of HRAs considerably reduced the time of experimental studies, reduced the cost of the unit, unified the unit manufacturing quality control, improved the prediction of its efficiency, and provided the maximal stability with respect to this load. The basic construction of the first-generation HRA in a heterogeneous NRE is shown in Fig. 2.1. In the upper part of HRA housing is placed an ordered set of ceramic elements that are not attached to each other. The lower part of HRA developed at the RDIET contains input channels of a working medium (hydrogen or nitrogen at a stage of cold purges).

The bench IVG-1 reactor, constructed to work out the HRA design for the NRE operation parameters [2], is a heterogeneous gas-cooled reactor with a water mod­erator and a beryllium deflector (Fig.2.2). It consists of stationary and removable parts. The stationary part includes housing 1 of the reactor with lid 2, deflector 7, barrels 3 for power control, biological protection units 6, and screens 8.

The removable part of the reactor core contains central assembly 9 with a set of 30 technological channels (TCs) 5 and central channel 4. The HRAs under study can be placed in both the TC group and the central channel, where the thermal neutron flux can be approximately doubled compared to its cross-section averaged value owing to a beryllium reflector surrounding the channel, which allows testing HRAs mounted in the central channel at forced (up to damaging) loads. The use of water in the bench NRE prototype instead of a hydride-zirconium moderator, which is close in its nuclear physical properties to zirconium hydride, expands the experimental possibilities of the reactor, allowing the replacement of units under study without constructive finishing, and improving the reliability of the reactor operation (Fig. 2.3).

Heat releasing elements are located in the HRA heating unit 29.7 mm in diameter and 600 mm in length (Fig. 2.4a). The heating section (HS) is divided into 6 heating parts, each of which contains 151 twisted-fuel elements and 12 semi-cylindrical fillers. Fuel elements and fillers are close-packed into a triangle lattice. The FEs is twisted along the axis. The relative diameter of the FE is 2.2 mm, the blade thickness is 1.24 mm, and the twisting step is 30 mm (Fig.2.4b).

The specific heat release and temperature in the FEs along the HRA length are distributed no uniformly, with a maximum at the central part of the HRA in the third heating section, while the temperature of the hydrogen medium monotonically increases toward the nozzle output (Fig. 2.5).

Heating sections are intended for heating the working substance up to a specified temperature. The first four HSs, counting from the entrance of the working medium into the HRA, consist of FEs made of a double solution of carbide-graphite’s; the last two of the HSs contain FEs made of a triple (ZrC + NbC + UC) solution. Each

section is assembled from FEs of three types containing uranium in different amounts and located in three zones. Profiling by the uranium load was performed to level the temperature field over the HS cross-section [4]. Formulas for calculation of principal stresses a1 and a2 is of cross-section double-blade fuel elements in a stationary regime were obtained by numerical method.

Value of numerical factor matches to conditions of heat exchange Bi = 0.25, value in brackets for Bi = 3.5 (involve a change range of these parameters); the Poisson’s ratio v is equal to 0.2, а-a coefficient of linear expansion, E—is Young modulus, Bi = atD/A, at is convective heat exchange factor, and D—the fuel element diameter.

The thermal stress changes essentially at high temperatures at the expense of stress relaxation. Thermal stresses relax especially strongly in fuel elements and casings. Calculations of a relaxation usually consider only an unsteady creep presented by experimental dependence: dex /dt = B(t)<rxm.

Remark It is seen, that tensile stresses Ф and azz, on a cooled surface of fuel element (Table2.1) are not equal and their values with maximum in two points

Table 2.1 Formulas for calculation of principal stresses ai and 02 is of cross-section double-blade fuel elements in a stationary regime

Cross-section form of Dangerous a1 a2

fuel element points

az = 0.022(0.019) Щ-D av = 0.015 0^

az = 0.027(0.022) aE%D 0

a* = °.°26

a* + a, = -0.014 а-*

B and C at the same thermal condition are lower than for the round rod of the same diameter. Really, for the round cylinder coefficient at a complex aEqvR2/ (A(1 — v)azz) is equal to 0.125, whereas azz for the double-blade rod in a point B is equal to 0.108 or 0.088, depending on criterion Bi, that is less accordingly on 13.6 or 29.6% than for the round cylinder.

For a one-dimensional problem connection; between deformation and stress with account of an elastic deformation and a creep strain becomes: dex/dt = B (t) axm + (1/E)dax/dt, where E—Young modulus.

The nonlinear thermal creep problem generally expressed by a numerical method. Calculation is made for the hollow ZrC cylinder which is heated up at a regime of linear increment of a heat release qv for 3sto20 W/mm3 with continuation of heating during 10 s at this constant power (Fig. 2.6). The properties data: heat conductivity, a coefficient of linear expansion, a modulus, and a speed of creep for calculation are given in Chap. 5.

At the initial stage of the HRA development, different types of FEs, in particular, spherical ones were considered [2, 3]. Spherical HRE systems have a high hydraulic resistance to the cooling gas flow, and therefore rod HREs were preferred. At the final stage, a double-blade twisted fuel element has been chosen from alternative assemblage of cylindrical elements with three and four blades (Fig.2.7a-c). For decrease of temperature stresses the fuel elements spliced of two, three, and more carbide wires were also offered (Fig.2.7d-e). The special place occupies a ball fuel element having minimum temperature stress, as it has no ribs, a covering on it are kept better then on other fuel elements designs. Cylindrical fuel element with temperature stresses takes the second place after the ball. It’s this property also led to the idea of creation “bladed elements”.

Design fuel elements sampling can be made by many criteria. For example, on the criteria of the least hydraulic resistance at the set factor of a convective heat exchange, on the greatest temperature of gas heating at the set maximum fuel element temperature, and also by the technological reasons.

Fuel elements, braded from two wires, with the same diameter, as the double­blade rod possess small advantage on temperature, and more heat release. However, application of the wire fuels was prevented by a high probability of their destruction already in the course of assembly operations owing to enough small strength of the seals between separate wires. As a result there was a danger of emersion of fragments of separate wires. Therefore, this modification was shut down and the double-blade twisted fuel elements was chosen with the worst thermal characteristics, but with the best fracture character not forming small fragments, in comparison with four blade fuel destruction [3].

The heat-transfer agent stream in the HGA with the radial current was formed by distributing and modular collecting channels (Fig.2.8).

The speed of heat-transfer agent in the radial direction of an order smaller value allows forming a heat reacting surface from the spherical elements of a submillimetric

Fig. 2.8 The HGA with a radial working medium flow

or millimetric range. Such ball elements possess the maximum thermal strength resistance. Their dimensions provide more developed surface heat removal and they are perspective for installations with the big thrust of 1,000kN. Problems of the thermal insulation of HGA casing disappear, since due to the radial hydrogen stream conditions realizing a wall cooling (by blowing of a refrigerating medium through a porous wall). The principal cause because of the radial circuit design of a working medium current yet had no further development, connected at a collecting irregularity of spheres placement, and a problem of flow maintenance of a given working medium along the HGA and absence of a reliable hot wall design of a modular collecting channel.

In the late 1970s and early 1980s RITP and RDIET have commenced an intensive work on developing a multimodal system [2, 4] capable of producing both jet thrust and electricity to power life support systems of the spacecraft. Besides the main nuclear propulsion mode, the NRE was to be operating at two generation modes: low-power mode for prolonged operation (several years), and high-power mode for 1/2 of the specified service life in the propulsion mode. The high power mode (HPM) presented no particular problems for the reactor. At low power mode (LPM) the heat transfer agent circulates only outside the NFA casing, while the heat from fuel rods is transferred to the casing by radiation through the thermal insulation. Such mode differs significantly from the propulsion mode, the former involving considerable temperature gradient across the NFA radius and uranium burnout (min. 3-5 %). Therefore, the applicability of the structural NFA parts and fuel rods under these conditions demands further research. First of all, the design and processing technology of the fuel rods should guarantee retention of fission products inside the rod for several years at temperatures of 2,000K under high vacuum or in H-containing working medium pressures of 0.1-0.2 bars.

Started almost five decades ago, the program for development of nuclear rocket engine (NRE) originally based on the political aims and priorities of conducting Cold War between the USSR and the US were suspended in the early 1990s due to the USSR having stopped funding of these works.

A heat-insulating packet (HIP) protects the housing from the thermal action of the working substance. Its constructive feature is a multilayer sectional packet structure that minimizes the possibility of penetrating cracks (to the housing) and allows varying the HIP material composition over both its length and its thickness. The outer casings of the HIP made of pyrolytic graphite provide, along with heat insulation, a ‘soft’ contact with the housing, thereby facilitating the assembling of the construction and minimizing the abrasive action of heat insulation on the housing. The inner casings are thin-wall carbide-graphite cylinders. In the low-temperature region, they are made of zirconium carbide, and in the high-temperature region, they are based on niobium carbide. These casings serve as the supporting frame of the HRA preventing the entry of fragments of heat-insulation elements into the channel of the heating sections. The cases ensure the assembling of HSs and their mounting into a heat — insulating packet and reduce the erosion and chemical action of the working substance flow on the heat insulation. Casings made of low-density pyrographite and porous zirconium and niobium carbides are placed between pyrographite casings and cases. Casings made of low-density pyrographite are located in the low-temperature region (T = 1,500-2,000K). At higher temperatures, casings made of the so-called foliation consisting of carbide layers in a graphite matrix were arranged in the first version of the construction. In the second version, they were replaced by casings made of porous zirconium and niobium carbides.

For the thermal flow density in the cooling channel up to qs ^ (2^2.5)MW/m2 and the maximum temperature on the surface of the inner case of the packet 3,000 K, the thermal insulation should ensure the temperature on the metal housing of the channel not more than 760 K, which means that the effective heat conduction of the packet should not exceed 3W/mK-1 at T = 1,500 K.

The temperature distribution over the heat-insulating packet thickness is deter­mined by four heat transfer mechanisms in the gaps: molecular, convection, radiation, and contact resistance. The heat transfer depends on the gas composition and pressure in the gaps, the gap width, the wall temperature, and the gap eccentricity between the walls [6]. Estimates made for two versions of the HIP design (Fig. 2.9) show that the gap between the walls of the casings has the strongest effect on the heat transfer.

A bearing nozzle unit (BNU) supports the HSs and partially supports the HIP. All the axial stresses produced by the pressure drop are transferred through this unit along the hot HRA channel to the housing. In addition, the BNU provides the ejection of the working medium with specified parameters. To minimize the effect of possible cracking, the BNU consists of sections. It contains a bearing grid (BG), a bearing socket (BS), and a nozzle unit. Bearing grids are in the form of a ‘sintered’ unit of four-blade rods made of solid ZrC and NbC carbide solutions of equimolar composition. To increase the bearing area and provide a cylindrical surface, segment facings are attached to the side surfaces of the rod unit. The bearing socket consists of three successively arranged inserts made of carbide-graphite with a carbide protective covering. The nozzle is made of a set of conical carbide-graphite inserts.

The input unit is intended to produce a uniform gas velocity field at the input to the HSs, face screening of the neutron flux, compensation for thermal expansions of HSs, HIP, and BNU, and tracing of pulsed tubes and thermocouples used for measuring the working substance parameters. The input unit contains springs for compensating temperature expansions, a pyrographite casing, a beryllium cup serving as a face

deflector, a gas inlet, and an input grid consisting of a high-pass grid and four grid rows.

The force elements of the HRA operating at high pressure drops are made of hydrogen-compatible materials having a high specific strength, high radiation resis­tance, and low hydrogen embrittlement, especially in the soldering and welding joints. The metal housing of the HRA in the IVG-1 reactor in the active core region was made of an AMG-5 aluminum alloy and of 18-10 steel in other regions. The HRA housing in the IR-100 reactor, which should be at the external pressure of the order of 10MPa during flight, has two layers. The inner layer consisting of beryl­lium inserts provided the housing stability, while the outer layer, a thin steel jacket, provided the HRA sealing.

Among the zirconium hydride moderators developed and proposed so far, the simplest moderator contains a vertical set of thirteen perforated zirconium hydride discs that are closely adjacent to each other and have the diameter equal to that of the reactor core and thickness 50 mm each. The discs have 37 holes 41 mm in diameter for the HRA and 372 holes 3 mm in diameter for the flow of cooling hydrogen, which provides the required temperature field in the discs. This construction offers a simple solution to the problem of profiling the cooling system of the moderator with the energy release up to 1MW-cm-3 (Fig.2.10).

The most thermal stressed part of a moderator disk is the average cross-section. The maximum temperature drop at design reactor condition makes ~200K. The stress-deformed condition of moderating material is defined at two-dimensional problem of thermal elasticity for a case of plane stress deformation. The maximum stress on a surface of cooling channels in diameter of 3 mm with surface temperature Ts = 475-513 K attains magnitude = 80.5MPa [8]. This figure does not pro­vide normal working capacity of moderator on a design condition. Therefore, other alternative of the moderator block with the reduced stresses has been offered.

Fig. 2.10 Neutron hydride moderator block

Besides a disc design, the circuit scheme of rod moderator from zirconium hydride was offered [2], more optimum from the point of view of thermal strength. Such moderator consisting of rods circular profile allows:

• to organize a profiling of an active zone of moderator and to increase efficiency of the reactor;

• to raise thermal strength resistance of the elements making a zone of moderating material at the expense of its size decrease;

• regular distribute energy of dynamic loads in the volume of active zone.

The nuclear reactor deflector block presents itself the hollow cylinder consisting of twelve sectors, each of which contains a compound drum with controlling rods. A rejecter material is a beryllium.

The stationary temperature condition in the deflector block is attained after т = 24 s. The maximum stress occurs at the moment т = 5 s and exceeds stationary stress on the average more, than in 3 times; their values at т = 5 s, a0Max = 50MPa, and a0min = —120 MPa.

The safety factor on ultimate strength members of the deflector makes 3 that formally meet demands of normative documents.

The basic parameters of the NRE are the temperature, the neutron flux, and the average level of pressure and pressure pulsations of the working substance in different regions of the HRA. A measuring system in the ground IVG prototype [9] provides a reliable control of the working process in the HRA and ensures autonomous emer­gency protection of the mount over the HRA parameters during tests. In each HRA, two measurements of the gas temperature were provided at the end of the third HS, one measurement of the gas pressure behind the BG and two measurements of the housing temperature in the middle of the HS. The working medium temperature in one cross-section of the HS in the HRA is controlled with two to five zone ther­moelectric converters (TECs). Zone TECs made of a tungsten-rhenium alloy are used to measure the inhomogeneous temperature distribution in the radial direc­tion. In general, inaccuracy of measurements is a function of several factors, such as thermal-physical properties of a junction, neutron fluence, temperature, velocity, and pressure of coolant flow. Complex of metrological studies [10-12] that were carried

Fig. 2.11 Input channels of a working medium and communication systems for measurement of pressure and neutron flux at lower part of IVG-1

Fig. 2.12 Test reactor RA capacity 0.5 MW: 1 the case in diameter of 586 mm and height 700 mm; 2 additional graphite deflector; 3 regulating drum; 4 moderator; 5 ampoules with fuel; 6 technolog­ical console; 7 adjusting mechanism; 8 casing; 9 a deflector

out made it possible to manufacture several thermocouple devices ensuring control of a thermal mode of the fuel assembly testing (Fig. 2.11).

Inaccuracy of measurements of console type thermoelectric transducer is 2.5%, while antenna type thermoelectric transducer has minimal inaccuracy of measure­ments which is not higher than 1.9%; at the same time, durability of the latter device during testing is several seconds only. Small-sized thermal-electric neutron detec­tors (TEND) with a diameter not more than 2 mm have been successfully used for detecting profiles of the thermal neutron flux density over the height and in different sections of the active zone of the NRE. TENDs do not require the external power supply, and they are not sensitive to a value of isolation resistance under conditions of reactor radiation. Endurance radiation tests confirmed long-term operability of TNDs up to the thermal neutron fluence of 2 x 1021 cm-2. Pressure and temperature of hydrogen at the FA outlet and at the jet inlet define major parameters of the engine, the thrust and specific impulse.

The working capacity of fuel elements for the multimode NRE at low power regimes (LPR) was investigated from the beginning of 1987 [9] in the ampoule design (RA) which was reconstructed from reactor IRGIT No 3, capable to work continu­ously for a long time (months) for development of various fuel element geometry and compositions (Fig. 2.12).