Radioactive changes in the moderator material, zirconium hydrate, during NRE oper­ation in the engine regime (1,000 s) are insignificant. In the power regime after the irradiation of zirconium hydride in a reactor by the integral neutron flux of the order of

Fig. 5.7 Relative length change of graphite samples of various grade: RBMK (7), VPG (2), MPG 6 (3), irradi­ated under neutron fluence at temperatures 1,570-1,770 K.

■ • A Samples are cutting per­pendicularly to compaction axis; □, О, Д Samples are cutting in parallel direction to compaction axes; 1, 2, 3 measurements after irradia­tion; 4 measurements during irradiation

1021 cm-2, hydride swelling by 1.5 % is possible [13]. The temperature dependence of the zirconium hydride swelling is nonmonotonic. The minimum of swelling is observed at 550K and at 800 K is the maximum. The electric resistance changed on 7 % after 3,000h of irradiation by a fast neutron flux with the density 3-1020 cm-2 at 855K [14]. The heat conduction of ZrH1,8 irradiated by a fast neutron flux with the density 3.2-1021 cm-2 at 320 K decreased from 31 to 18Wm-K-1 [14]. As awhole, the results of irradiative studies suggest that the efficiency of the HRA in the NRE in the engine and energy regimes is retained.

This book covers the problems of theory and practice of developing core elements for a high-temperature Nuclear Rocket Engine reactor (NRER). As a concept, the problem is solved through a non-separable interrelation of the structure, properties and performance patterns of the structural elements. Due to high brittleness of materials used in a structure, on the basis of interstitial phases, i. e., carbides, nitrides, hydrides, special focus is put onto reasoning stability of the core elements being exposed to high heat and neutron flux, high stresses, aggressive gases, and ultra-high temperature. Such a variety of performance parameters required development of a whole range of new pre-reactor and reactor methods of material testing and execution of a complex investigation of the basics for the structure formation of these materials, in order to ensure the maximal reliability of the structures operation. Development of Fuel Elements (FE), simple by their shape, but operating under extremely heavy conditions, required involvement of the experts in various sciences, such as physics, mechanics, heat engineering, reactor and radiation material science, production technologies of nuclear fuel, and structural materials. The developed phenomenological and statistical-thermody­namic theories of three-component interstitial alloys and the set of laws for modification of diffusion-controllable processes (creep, sintering, carbonizations, and corrosion) have deepened present scientific presentation of these processes and have got wide recognition in the scientific community. Doubtless interest is pre­sented also on irradiation behavior of carbide materials and bearing capacity of fuel and constructional materials on the basis of interstitial phases in conditions of an irradiation neutron fluxes.

There are numerous separate works discussing development and production of FEs for thermal neutron and fast neutron reactors, but there is no book covering a problem of developing core elements for Fuel Assemblies (FA) integrally. The fact, that results of exploring on the NRE reactor are disseminated under various publications of periodic journals or express collections, access to which is com­plicated and own experience are great, so it became the incentive reason for writing this monograph in which data on a resolving of a problem of FA creation for high-temperature solid phase NRE reactor, characteristics of materials and

principles of designing are generalized. Author realizes that the frame of con­cerned subjects in a book is restricted. Many problems of design, neutron-physical and hydro dynamical characteristics of the NRE reactor are handled in the book Nuclear rocket engines [1].

Many scientific investigators and engineer’s of the Research Institute of Pro­duction Research Association (RIPRA) ‘‘Luch’’ participated in a development, manufacturing and material tests of the NRE are presented in this book [2]. The circle of authors shared in creation of the NRE is rather wide and, certainly, is not terminated by the authors presented in the bibliographic list of this book. Many actively worked authors who died, did not publish in due time their manuscripts in the open printing. So it should note the big contribution of Ph. D. N. Poltoratsky and Doctor of Science L. B. Nezhevenko in the development of the fuel elements technology HRA from carbide and carbide-graphite materials and Ph. D. A. L. Epstein for the material science organization and technological works on neutron moderator from hydrides of zirconium and yttrium. The author remembers the creative contacts with them with gratitude, and also a fruitful cooperation with many workers with whom he had the luck to work at the NRE elaboration in scientific research institute of RIPRA ‘‘LUCH’’ since 1962 year.

The author expresses gratitude to Professor R. A. Andrievsky for the review of the manuscript and useful advices and to engineer A. P. Abramova for the book formatting.

St. Petersburg Anatoly Lanin

References

1. Demjanko, J. G., Konuhov, G. B., Koroteev, A. S., Kuzmin E. P.,& Pavelev, A. A. (2001). Nuclear rocket engines (p. 413). Moscow: Open Company ‘‘Norm — inform’’.

2. RIPRA ‘‘Luch’’. (2004). Affairs and people. Podolsk. RIPRA ‘‘LUCH’’. Ed. Fedik I. I. Podolsk, 455 p.

Furnaces developed and constructed at the institute for sintering and consolidating powder samples and deposition of coatings on products at high temperatures (up to 2,500-2,700K), which were unique for that time, were used to prepare samples required for studies and, later, components of the NRE core.

The general ceramics processing technology involves the following steps [14]:

1. Synthesis of powders with a given chemical composition, phase composition and grain size distribution;

2. Preparation of powders for forming;

3. Green part forming;

4. High-temperature consolidation;

5. Process quality.

Sintering is a crucial stage for structure evolution. Sintering is a thermally activated process (spontaneous or involving the application of external force) which features transformation of a system of contacting solids or porous media into thermodynam­ically more stable state through minimization of the free surface area. The driving

Fig. 4.4 Installation for hot pressing (IHP-4) with pressing force up to 200tons and working space 900 x 1800 mm

force for this transformation is the excess free energy. Sintering is evident both as a dimensional change (mostly as shrinkage), and as significant change in structure and properties, both approaching these of compact materials. At the RIHRE “Luch”, there were developed special electrothermal vacuum systems for synthesis, sinter­ing, and hot pressing of refractory carbides, and systems for synthesizing fine carbide powders (Figs.4.3 and 4.4).

Details of kinetics of density, structure, and properties evolution during zirconium carbide sintering are described [15, 16]. ZrC powders containing 87.8 % Zr, 11.4 % of Ctotai, 0.15 % of Cfree, 0.018 % N, 0.62 % O were synthesized by carbon-thermal reduction at T = 2,100 °C. The synthesized powders were ball-milled in a vessel lined with refractory ZrC plates. Average particle diameter of powders was 4.5­5 ^, with BET surface area of 4.35m2/g. Green parts were formed by thermoplastic extrusion, dewaxed in a vacuum furnace at final temperature of 350 °C with heating rate of 10K/h, and dwell time of 2h. Final sintering was carried out in the graphite

heater furnace at 1,400-2,800°C with dwell time of 0, 15, 30, 60, and 120min; heating and cooling rates were 800-900K/h (Fig.4.5).

High values of Y, a and electrical conductivity were obtained at 2,200-2,400 °C. Sintering in hydrogen and argon atmospheres yielded similar results, while vacuum sintering yielded slightly poorer properties, which is attributable to greater gas evo­lution during sintering possibly leading to more porous compact structure. Similar results were observed during sintering of ZrC+NbC and ZrC+NbC+UC systems [11]. The further increase of sintering temperature to 2,600-2,800 °C did not lead to better density or strength of the sintered samples. Poorer sample consolidation could be due to the effect of gases in the closed pores. At the overall porosity of ~6-7 %, the majority of pores are closed which greatly hinders further consolidation.

Higher porosity of samples sintered at 2,800 °C may be attributed both to gas pressure and to coalescence of gas-filled pores. Lesser strength of samples sintered at 2,600-2,800 °C is due not only to poorer density and significant grain coars­ening, but also to pore segregation at grain boundaries leading to formation of the

Fig. 4.6 Features of micro­scopic ZrC structure sintered at different temperatures and various medium [16]. a Vacuum, Ts = 2,000 °C,

T = 15 min x1000. b Vacuum, Ts = 2,200 °C,

T = 15min x1000. c Vacuum, Ts = 2,400 °C,

T = 15 min x1000. d Vacuum, Ts = 2,600 °C,

T = 15 min x1000. e Hydro­gen, Ts = 2,700 ° C, t = 5min x 1000. f Argon, Ts = 2,700 °C, t = 5min x 1000

highly specific crack patterns (Fig. 4.6). Samples sintered at T > 2,600 °C show pre­dominantly intergranular fracture, while those sintered at lower temperatures show mostly transgranular fracture. The relative density and strength a of fuel rods based on ZrC+UC and sintered at 2,200 °C in hydrogen reach 99 % and 650 MPa, respec­tively, while vacuum sintering at the same temperature yields values of 88% and 360 MPa, respectively [11].

When analyzing the sintering process one should consider porosity features. At relative density nearing 90 % of TD, the residual porosity is nearly always an open one, i. e., pores are connected to a surface. At overall porosity of 7-8 %, all pores became closed. As temperature rises or as sintering time increases, the powder compact becomes denser and the shrinkage increases, possibly reaching 20-25%. Simultaneously, there is a change in properties: density change could be traced by monitoring the decrease of the electrical resistance.

During sintering in gas atmosphere at 2,000-2,400 °C, and in vacuum at 2,300­2,500 °C densification mainly proceeds by decreasing the closed porosity; further temperature rise to 2,600 °C yields virtually no change in closed porosity. At this stage, pores coalesce and become spherical (see Fig. 4.6c, d). The number of pores decreases, while their size increases. At sintering temperatures above 2,600 °C, the closed porosity and overalls porosity values grow. Sometimes surface cracks appear on samples, leading to decrease of the closed porosity. This particular phenomenon was most frequently observed in samples sintered in argon at temperatures above 2,700 °C.

Bubble-shaped pores at this stage of sintering are pinned by boundaries at which these pores tend to segregate and coalesce (see Fig.4.6e, f). The shape of the pores changes, vast cavities appear, and part of bubble pores migrates to the sample surface as a result of coalescence of two gas filled pores of equal radius; the volume of a newly created pore may increase (provided that the volume if gas before and after the coalescence remains constant), which leads to the swelling of samples. In ZrC samples sintered at temperatures above 2,600 °C, the volume increase and density drop were observed. In some samples, this phenomenon was observed already at sintering temperature of 2,550-2600 °C, especially in argon atmosphere. The results of analysis of the furnace chamber atmosphere show that during sintering in Ar at pore closing stage (T at 2,000-2,200 °C) the major impurity is CO which may become trapped in the closed pores of the sample. Moreover, CO also forms during sintering through the reaction of free carbon with oxygen, and during rapid heating at 2,000-2,200 °C this CO may also become trapped in pores.

Samples with the green density p of 52-57 % were sintered in vacuum at tem­peratures of 1,000-2,250 °C, and subsequently in Ar atmosphere at Tmax= 2,500­2,700 °C. The shrinkage kinetics during sintering in a furnace with the special fixture for controlling the linear dimensions of the sample was constantly measured with an accuracy of ±10 ^.

The sintering behavior of the system “ZrC+ 2^10 % C (diamond) up to the tem­perature of 1600-1800 ° C was similar to that of the pure ZrC [17]. In the temperature range of 1,650-1750 °C corresponding to the maximum volume change of the dia­mond (due to its polymorphic transformation), the shrinkage rate dropped nearly to zero. Further temperature increase for this composition leads to significantly accel­erated densification. At this stage, the sample shrinkage curve of the ZrC-2.5 % C system is similar to the one of pure ZrC. Changing the diamond particle size from 50 to 1 ^ leads to a change in shrinkage curves. The absolute value of the density loss (caused by polymorphic transformation of diamond) decreases. The sintered compositions being studied show predominantly aggregated structure with carbon phase particles located primarily on the grain boundaries of the carbide matrix.

The general regularity of structure formation of ZrC, ZrC-C, NbC-C is the high speed recrystallization at sintering temperature above 2,250 °C: The grain sizes of compositions increase to 10-20 ^ with growth of carbon phase after sintering at 2,500 °C. Below this sintering temperature, the grain size of sintered composition are comparable with initial sizes of carbon corpuscles (5 and 2.5 ^).

The pores placed on boundary lines of grains, suppress grain growth at sinter­ing. An average grain size of carbides increases slightly (from 20 to 25 ^) at the sintering temperature interval 2,500-2,700 °C. As a whole the structures of sintered compositions ZrC, NbC with carbon additives (diamond, graphite) is characterized by the uniform distribution of pores in a matrix [18]. Estimations of the sintering energy activations Q of compositions ZrC-C, NbC-C show that Q in the investi­gated interval of temperatures is changeable [17]. At temperatures 1,300-1,400 °C Q = 65-80 kkal/g-form for compositions ZrC-C and NbC-C Q = 70-75 kkal/g — forms. At higher sintering temperature Q increases to 115-125kkal/g-forms for (ZrC-C) and to 120-135 kkal/g-forms for (NbC-C). The received values Q at

sintering of the compositions ZrC-C, NbC-C proceeding from the parameters of a self-diffusion component in carbides ZrC, NbC [2], testify to predominance of the boundary diffusion contribution in mass transfer at sintering to 1,300-1,400 °C. Role of the volume diffusion controlling the shrinkage at sintering becomes appreciable at temperature growth.

It is known that the temperature pattern formed in a powder pressing at heating­cooling can essentially influence on the character and intensity of sintering. However, the data on an agency of heating and cooling speeds on a process of carbide materials sintering in the literature are restricted by consideration of small speeds of heating­cooling near 500-1,000K/h. In this connection, an undertaken attempt [19] to use high-speed heating-cooling when temperature-kinetic parameters of heat treatment can become crucial. High temperature heating up to 2800 K is used for the carbide pressings ZrC0.95 in diameter of 2.5 mm with a speed 600K/s allows for increment sharply to 103 times speed of a shrinkage and to raise an initial density from 71 to 98% (Fig. 4.7).

The influence of sintering parameters on the sintered ZrC structure and properties are showed on Figs.4.8 and 4.9 accordingly. Sintering of ZrC samples by rapid nonisothermal heating takes place in a complicated manner. In the first stage of the sample heating to 1,300 K, for the first 2 s thermal elastic stresses of the order 30- 40MPa (Fig. 4.7) are developed and are capable of creating local plastic deformation of the material on rough surfaces of the particles. The presence of this phenomenon is supported by an experimental study of the dislocation start from the indentor print on the single crystal ZrC0.9 performed at a temperature of 290 K and subsequent heating to temperatures 600 and 850 K. The initial stresses as for dislocation movement at these temperatures are at the limit of 5-8 and 2-3 MPa, respectively [20]. It should be noted that local stresses on roughnesses with the size of hundreds of thousands of angstroms can exceed average-mass thermal elastic stresses by one to two orders. And the latter can relax at the very initial stages of sintering.

In the second stage, which occurred with further temperature increase up to 2,000 K, the shrinkage, perhaps, is defined by mutual slip of particles as a whole

along the boundaries between them. They move into the voids and to the free surface of the sample (Fig.4.8a-d) according to the mechanism proposed in [19, 21].

After overcoming the resistance of retainers on the definite part of the surface of the particles and their ‘activation’ the boundary diffusion viscosity becomes low and shrinkage sharply increases. Random boundaries that are not connected with initial boundaries between particles are formed. Shrinkage rate decreases with porosity decrease.

At the third stage of heating to temperatures higher than 2,200 K in the central part of the sample, the final stage, coagulation of pores takes place on far from the free surface, isolated pores appear. Further densification takes place due to decrease of the peripheral porous layer with open pores and a reduction in the number and volume nonisothermal sintering of the samples at mainly of the isolated pores (Fig. 4.8d).

An appropriate choice of the nonisothermal heating parameters can provide a fine — grain regular-pore structure (Fig. 4.8b), with high level of strength at appreciable time decrease of sintering.

An important factor determining the operation conditions of HRAs is the corrosion resistance of materials in hydrogen with methane added [1, 2].

The extent of carbides interacting with gases is defined by gas activity with carbide components and gas partial pressures in a mix [3-5]. The maximum activity is the oxygen, which interact even at low partial pressures (10-3-10-1vol%) harder above than hydrogen; therefore, the interaction with oxygen will be observed first. The mechanism of ZrC interacting with oxygen can be presented by following equations (Table 6.1). The detailed oxidation mechanism description of the NbC and solid solutions ZrC-NbC is complicated by two reasons: first, there is a row of the flying oxides in the system Nb-O which is sublimated at high temperatures; second, the oxidation of these carbides was studied in rather narrow temperature intervals.

Kinetic characteristics of oxidation of ZrC and NbC are resulted in Figs.6.1 and 6.2.

Determination of kinetic regularity in an intermediate interval of temperatures 1000-1300K is complicated because of an oxide film fall and an unknown surface of oxidation. The resulted data on oxidation kinetics refer to the materials practically with a stoichiometric composition. Decrease of C/Me to 0.9-0.8 in all carbide mate­rials reduces reaction rate constant. High-temperature oxidation (2,100-2,500K) in hydrogen (H2 + 0.07 ^ 0.1%O2) leads to decrease in strength of ZrC approxi­mately two times for the first 1,000 s of interacting (Fig.6.3). At higher tempera­tures (>2,600K) in the specified gas mix ZrC samples are fractured after 500 s of interaction. A source of rupture is a formation of metal zirconium on reaction 5 (Table 6.1). Metallographic analysis of a fractured places after the sample exposure at T = 2,700 K showed the segregations of metal zirconium in an oxide film.

The interaction of carbides with hydrogen at high temperatures is accompanied by structural changes, the formation of hydrocarbons, and a decrease in the carbide stochiometric according to the reaction

MeC + y/2 H2 = MeC + CxHy.

A. Lanin, Nuclear Rocket Engine Reactor, Springer Series in Materials Science 170, DOI: 10.1007/978-3-642-32430-7_6, © Springer-Verlag Berlin Heidelberg 2013

Fig. 6.1 Oxidation of zirconium carbide in air. (Z1C0.95, Ctotal = 11-4%. Cfree = 0.3; O + N = 0.10%;

= 3-5%; L = 8 ц) [4]. 1—T = 850 K, 2—T =

900 K, 3—T = 950 K,

4—T = 1,000 K, 5—T = 1,050 K

Change of lattice parameters of ZrC after an exposure time in hydrogen allows sizing up the change of relationship C/Zr on cross-section of the sample at the expense of a carbon removal (Fig. 6.4). Calculated estimations are coordinated qualitatively with the chemical analysis data.

The type of the interaction of UC-ZrC, UC-NbC, and UC-ZrC-NbC systems with hydrogen is similar. Only the degrees of decarbonation, the carbon concentration gra­dients over a sample cross-section, and recrystallization temperatures are different. This suggests that all the systems under study have the same mechanism of interac­tion with hydrogen. On the surface of samples, the reaction of carbon with hydrogen proceeds with the formation of CH4 and C2H2 and the evaporation of metal atoms, while the diffusion of carbon from the center to the surface and recrystallization occur in the bulk. Calculating the interaction for the scheme of the process under study requires solving nonstationary diffusion problems with the boundary condi­tions simulating a mass exchange with the working substance in the presence of heterogeneous chemical reactions (Fig. 6.5).

The total rate of a material removal for ZrC and NbC is maximum at first moments of an exposure time and the temperature dependence is expressed by the Eqs. (6.7) and (6.8):

The total entrainment rates for double ZrC-UC and triple ZrC-NbC-UC solid solutions in hydrogen-methane media with 0.656 volume% of CH4 at 3,150 K are approximately the same (0.94-10-6gcm-2 s-1). High temperatures in experiments lead not only to a change in the chemical composition of materials but also to a change in the density of carbides and to an increase in the grain size. The appearance of a carbon concentration gradient over the sample cross-section at temperatures up to 2,500 K for 1,000 s leads to the formation of compressing stresses up to 500MPa and to an increase in the strength due to a change in the lattice parameter.

The interaction of carbides with hydrogen at 3,100K for 1,000s produces an inhomogeneous concentration of carbon not only but also causes a change in the structure, a decrease in the density, and a reduction in the strength to 50 % (Fig. 6.6). It should be noted that exposure at a lower temperature near 2,600K may even increase strength due to the formation of residual compressive stresses. The results of model tests agree well with the HRA tests in the IVG-1 reactor.

Decarburization of NbC in hydrogen occurs more intensive. Transformation of NbC into Nb2C occurs at 2,450K after 6,000 s. The gradient of concentration of carbon on cross-section is insignificant though Nb2C grains are observed all over

cross-section. Active growth of grain is observed at a surface of the sample since 2,500K. Temperature lifting to 3,000K leads to intensive decarburization and the damage of samples (Fig. 6.7).

The general entrainment of niobium carbide is connected with carbon loss, and the metal component of carbide evaporates slightly, as opposed to ZrC when an apprecia­ble evaporation of zirconium is observed. This fact supports higher concentration of
carbon in ZrC after interacting. Interaction of solid solutions ZrC-NbC with hydro­gen at high temperatures along with decarburization, the enrichment of surface by more refractory component of niobium is observed.

Apart from the chemical action of a high rate working substance flowing around an element, a force erosion action of the flow on the surface is also possible; as a result, some weakly attached carbide particles are carried away by the flow.

Based on the dependences of the entrainment rate of the carbon component from carbides, another protection method was proposed in which hydrocarbons are added to hydrogen in the amount at which their mean concentration in the flow becomes equal to their equilibrium concentration over the surface. Then the carbon entrain­ment must be zero. But the equilibrium composition of hydrogen can be achieved not over the entire HRA length but at individual points only. In other places, either the entrainment or deposition of carbon on the washed surface occurs. Therefore, it is reasonable to regard the method of protective additions as supplementary to the main method of protection with the help of coatings.

Thermal stresses, a1, in a coating of thickness h1 originate even at uniform heating to temperature T due to a difference of coefficients of linear expansion of a coating a1 and substrate a2. Accepting the condition of equal deformation on the boundary line between coating and substrate, the stress in the coating can be expressed:

_ E1E2 a — a2) T

ai ~ IE1 (1 — 1×2) + E2(1 — 1×1)

where index 1 refers to coating and index 2 refers to substrate. E is modulus elasticity, ц is Poisson’s coefficient. Resistance of coating can be determined by the level of limiting tensile stresses.

The thermal stress destruction of a coating can occur as a result of coating cracking or at the expense of its peeling from a substrate to the subsequent destruction of a brittle thin covering. The fracture aspect is defined by extent of adhesion and parameters of a thermal loading (Fig. 6.8).

The corrosion resistance of coating made of low-density pyrographite (LDP) (1.35-1.6 g cm-3) in the HIP, their weight in the hydrogen flow at 1,570K decreases for 6,000 s by approximately 3-8 %, while the strength decreases by 30 %.

The strength of casing from ZrC composition with pyrographite after annealing essentially decreased. Some casings (No 2, 5, 6, Table6.2) were destroyed at a light touch.

The loss in the mass of cases made of PGV pyrographite with distinctly anisotropic properties (density from 1.2 to 2.25 gcm-3) in hydrogen at 2,300 K for 4,000 s reaches 45 %. Due to a low porosity, the interaction of uncoated pyrolytic graphite with hydrogen, for example, at 2,100 K is two times weaker than that for usual pressed graphite types.

The high pace of development and the high scientific and technological level of atomic power engineering achieved in the USSR in the second half of the 1950s created objective prerequisites for the construction of fundamentally new types of stationary and mobile nuclear energy reactors and facilities intended, in particular, for cosmic purposes, such as nuclear rocket engines (NREs) and small-size nuclear- electric energy converters [1, 2].

The development of NREs in the USSR was initiated in 1955 when I. V. Kur­chatov, S. P. Korolev, and M. V. Keldysh met to discuss the possibility of creating a rocket with an atomic engine to enhance the defensive power of the country in response to the Rover program for developing NREs, started in the USA (Fig. 1.1).

In 1956 and 1958, two government acts for the development of work with the aim of building nuclear rocket engines were signed in the USSR [2]. At this stage, preliminary design work was started to create a mounting base for testing NREs, and material technology investigations were initiated. The heat exchange and hydrody­namic conditions were studied at Research Institute-1 [currently, Research Institute of Thermal Processes by V. M. Ievlev (RITP)]. Studies in the field of neutron physics and reactor control were performed at the Obninsk Physical Energy Institute (PEI) and at the Kurchatov’ Atomic Energy Institute (AEI). The technology of refractory materials and heat-releasing elements (HREs) was developed at Research Institute-9 (currently, the A. A. Bochvar All-Union Research Institute of Inorganic Materials, ARIIM). The problems of selecting materials for NREs based on carbides of tran­sition metals and graphite in gas media were also partially investigated, beginning from the 1960s, at the All-Union Institute of Aviation Materials (AIAM), the State Institute of Applied Chemistry (SIAC), the Institute of High Temperatures (IHT), the Graphite Research Institute, and the Ceramics Department of the Leningrad Tech­nological Institute.

In 1957, researchers at the RITP proposed a principle of refining individual ele­ments of units of the NRE reactor on electrothermal and plasmatron mounts, which reduced the extent of reactor tests. Different NRE schemes were considered. The sim­plest NRE with a reactor with a solid-state core (named the A scheme) can produce

A. Lanin, Nuclear Rocket Engine Reactor, Springer Series in Materials Science 170, DOI: 10.1007/978-3-642-32430-7_1, © Springer-Verlag Berlin Heidelberg 2013

a thrust 450 s. The NRE with a gas-phase reactor (the B scheme) could produce a specific thrust up to 2,000 s, but problems to be solved in its construction, namely, the confinement of a uranium plasma with the temperature up walls of a heat-releasing assembly (HRA) from the plasma, were much more severe than those in the con­struction of the A scheme. It was decided to begin the development of the A scheme (while the work related to the B scheme was continued as a research study).

Within 2 years, two teams were actively involved in the development of the first NREs. A ground-based IVG-1 (research high-temperature gas-cooled) reactor, a pro­totype of an in-flight NRE version, was developed at the Research and Development Institute of Energy Technology (RDIET) headed by N. A. Dollezhal’ and at AEI under the supervision of N. N. Ponomarev-Stepnoi. The second team developing a ground-based IR-100 engine was headed by V. M. Ievlev, the scientific chief of RITP, A. D. Konopatov, the chief designer of the engine at the Development Labo­ratory of Chemical Automation (DLCA), Voronezh, and V. Ya. Pupko, the scientific chief of the reactor at the Physical Energy Institute (PEI), Obninsk. The first carbide- based HREs were proposed and manufactured at Research Institute-9. The expansion of NRE studies required the manufacture of a great number of HREs for experi­mental HRAs. In August 1962, the Ministry of Medium Machine Building of the USSR decided to create the Research Institute of Heat Releasing Elements (RIHRE) [currently, the “Luch” Research and Production Association (RPA), a Federal State Unitary Enterprise (FSUE)] with an experimental plant providing technological pos­sibilities for the rapid complex development and production of new types of nuclear fuel and HREs [2]. This resolution was preceded by the very important decision of 30 December 1959 about the inexpediency of further developments of military nuclear ballistic rockets (due to progress in the development of chemical-fuel engines) and the necessity of continuing NRE developments for space launchers [1].

The basic challenge encountered by the developers of NREs was to prevent dam­age to the ceramic active core and, most of all, HREs caused by thermal stresses. It is known that thermal stresses are proportional to the product of the energy-release density, the elastic modulus of the material, the linear expansion coefficient, and the square of the characteristic transverse size of the construction. The energy release cannot be strongly reduced because such a reduction would result in an increase in the reactor weight with protection and would reduce the NRE advantages to a minimum. Reducing the HRE characteristic size can be done efficiently, but this is restricted by the fact that elements of very small sizes would be severely damaged by vibrational stresses, which are significant in rocket engines. In this case, the lim­iting characteristic transverse size of HREs is 2-3 mm. An alternative to the carbide HRE is the graphite HRE. Graphite has a unique thermal strength because its elastic modulus is almost two orders of magnitude smaller than that of carbides. However, a substantial disadvantage of graphite is that it quite actively interacts with hydrogen (everything comes at a price).

Unlike American scientists, Soviet researchers began to develop the NRE core using HREs made not of graphite, which is thermally stable but unstable in the hydrogen medium, but of carbides, which are brittle but are more stable in hydrogen [2]. While in many other scientific fields Soviet scientists had to ‘overtake’ foreign investigations, the NRE was developed without blindly copying ‘foreign’ samples, the experience of predecessors being critically analyzed. Soviet researchers decided (correctly, as the result showed) that protecting HREs from interaction with hydrogen is more difficult than providing its acceptable thermal strength resistance. Although graphite HREs was also developed in the USSR, they were regarded as a backup.

In 1962-1969, under the supervision of M. V. Yakutovich, director of the RIPRA, the material technology, technological, theoretical, and test departments were created to solve the following problems [2]:

• selection of the fuel and construction materials for NREs;

• development of the technological foundations for manufacturing devices from these materials;

• construction and computational and experimental demonstration of the efficiency of elements of the active core; preparation of project documentation and organi­zation of semicommercial production.

These problems were successfully solved by using a systematic and complex approach. Detailed theoretical estimates of the operational conditions of the devices, the required technological studies of the properties of materials, and the working out of technological versions were realized in the form of technological chains developed together with the experimental plant. In turn, experimental and test samples of devices were subjected to thermal, hydraulic, and resource tests in the IVG reactor located on the nuclear proving ground “BAIKAL-1” in the Semipalatinsk region (town Semipalatinsk-21, later Kurchatov, Kazakhstan). The working area “BAIKAL-1”

Fig. 1.2 Director of the RIPRA “LUCH”, Doctor of Science, laureate of many state prize M. V. Yakutovich (10.08.1902-29.06.1988)

is situated at 70 km from inhabited town Kurchatov and at 150 km from the Semi — palatinsk town (Figs. 1.2-1.4).

This paper is devoted to the history of material technology developments for manufacturing elements for the NRE reactor core at Research institute of Production Research Association [RIPRA] Federal State Unitary Enterprise (FSUE) “Luch” from 1962 to 1991. Because the results of NRE studies are scattered over numerous publications in periodicals or special collections of papers that are not easily acces­sible, we decided to generalize the data on the characteristics of the materials and the efficiency of elements of the NRE core in this paper.

Clearly, the scope of problems considered here is limited. For example, we do not discuss issues concerning the design and the radiation and thermal protection of the nuclear engine, or the neutron-physical and hydrodynamic characteristics of the reactor. These issues were partially considered in the book Nuclear Rocket Engines published in 2001 [1].

Because of the high brittleness of the materials based on interstitial phases (car­bides and hydrides), which are used in construction, special attention is paid in manufacturing active-core elements that are stable under the action of high thermal and neutron fluxes, stresses, aggressive gas media, and ultrahigh temperatures. Such a variety of operating parameters required the development of a number of new prere­actor physic-mechanical methods for testing materials at the RIPRA “Luch”, where diffusion-controlled processes (creep, corrosion, and radiation) were investigated in fuel and construction materials based on interstitial phases.

Beginning from 1976, HRAs developed for NREs were tested in the IVG-1 reactor, which was used to work out the elements of the reactor core. Later, a mount was constructed for testing an ‘engine’ version of the IR-100 reactor, which was tested for several years at different powers and was then converted into a material technology research low-power reactor, which is still working successfully till date.

In the late 1970s and early 1980s RITP and RDIET have commenced an intensive work on developing a multimodal system [3] 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 half 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 fluids at pressures of 0.1-0.2bars.

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.

The main output of the Soviet nuclear rocket engine (NRE) program from the ini­tial 60x years becomes the technological and experimental possibility of the ceramic active zone creation of the NRE for hydrogen heating at the maximum parame­ters values: A hydrogen reheat at temperature up to 3,100K; power fuel density to 35MW/l; duration of tests to 4,000 s; the maximum heating/cooling rates of fuel elements to 400/1,000K/s

The results of numerous studies performed at the RIPRA “Luch” for many years allowed constructing a scientific system of knowledge about the nature of refrac­tory compounds, mechanisms of their deformation and damage, and the features of their behavior in construction and fuel hardware during variation of the opera­tional parameters of reactors. Many developments at the RIPRA “Luch” have been acknowledged at international conferences, symposia, and exhibitions. A group of scientists has grown who have defended more than 30 Doctor of Science and 200 Ph. D. dissertations and published a few dozen monographs and a few hundred scien­tific papers on high-temperature materials and their applications. Some researchers became laureates of State Prizes of the USSR and Government Prizes and received titles of Honored Scientist, Honored Technologist, and Honored Inventor.

The workings out of NRE begun almost half a century ago caused, in due time, by political problems and priorities in conducting “cold war” between the USSR and the USA were suspended in the early 1990s in connection with the financing termination in the USSR.

References

1. Demjanko, J. G., Konuhov, G. B., Koroteev, A. S., Kuzmin E. P., & Pavelev, A. A. (2001). Nuclear rocket engines (p. 413). Moscow: Open Company “Norm-inform”.

2. RIPRA “Luch”. (2004). Affairs and people. Podolsk. RIPRA “LUCH”. Ed. Fedik 1.1. Podolsk, 455 p.

3. Ponomarev-Stepnoy, N. N. (1993). Creation history of NRER in the USSR (pp. 3-18). Third Branch Conferences “Nuclear Power in Space”.

4.3.1 Strength of Materials at Different Loading Mode

The change of loading mode from tension to bending and compression for ZrC, NbC, ZrC+NbC at the deformation rate 10-3 s-1 decreases brittle-ductile transition temperature Tb-d and rises the level of ductile deformation. The strength level of ceramics in brittle condition in most cases depends on the stress state considerably so while passing from tension to bending and compression the strength increases in relation 1:(1.5-2.0):(8-10) (Table4.4) [14]. This phenomenon is associated with kinetic peculiarities of crack propagation. Start of a crack in tension begins under attainment of critical coefficient of stress intensity Kic, which gives rise to further avalanche crack propagation till full body fragmentation. The crack initiated under compression at Kic is capable to grow uniformly under increase of continuous load only on a curved trajectory and tends to take its orientation toward the compression axis (Fig.4.10).

Eventually, the transition of equilibrium crack propagation into avalanche crack stage, leading to full fragmentation is made possible after interaction of equilibrium growth of cracks at loads many times higher than the initial load for crack start. Experiments and theoretical analysis [6, 14] show that the total body fragmentation arises under the combined development of the interacting cracks after some equi­librium growth, as the growth of single crack cannot cause fracture even at infinite load. The steady creep rate of carbides does depend on kind of loading (compression, bending, tension) at temperatures T > (0.65-0.70) Tm [22].

The variation of loading conditions radically altering the strength level has no influence on the variation coefficient = S/o and Weibull coefficient m (Table4.4), where S is a root mean square strength deviation and am is a mean arithmetic strength value. The W and m are in the range 20 % and 3-5, respectively, for monophase dense and porous ceramic. The W value increases by two times for heterophase carbides with carbon inclusions owing to damage of carbide matrix. It is significant that the strength variation parameters of various ceramics in the brittle state are not affected markedly by electron band structure and atomic bonding but are governed primarily by the variation of surface and volume flaws.

Transition from axial compression to multiaxial compressive loading by intrusion of indenter initiates the ductile deformation in carbides of transition metals even at 80 K. Temperature dependence of microhardness reveals deformation peculiarities (Fig.4.11), undetectable under other kinds of loading.

The variation of loading conditions radically altering the strength level has no influence on the variation coefficient W = Aa/a, and Weibull coefficient m (Table4.4), where A a is a root mean square strength deviation and am is a mean arithmetic strength value. Between distribution parameters m and W there is a corre­lation. Comparative tests of large and small sets of samples have shown that reliable enough estimation am can be obtained from a test of 5-7 samples, and distribution parameters an, W and m on 25-30 samples. The W and m are in the range of 20 % and 3-5, respectively, for monophase dense and porous ceramics. The W value increases by double for heterophase carbides with carbon inclusions, owing to damage of carbide matrix.

Transition from axial compression to multiaxial compressive loading by intrusion of indenter initiates the ductile deformation in carbides of transition metals even at 80 K. Temperature dependence of microhardness reveals deformation peculiarities, undetectable under other kinds of loading.

The first bend on the curve of microhardness takes into account the transition of gliding system {110} <110> to the system {111} <100> (Fig.4.11). This is typi­cal for transition metal carbides [20, 21]. The second bend associated with further development of ductile deformation and disappearance of cracks near indentations. The variation of load on indenter from 200 to 1,000g changes the temperature of crack (c) disappearance from 800 to 1,400K. The dislocation configuration around indentation (d) and elastic distortion zone (A), measured by Berg-Barret method, changing with orientation (Fig. 4.12), give valuable information about evolution of the deformation mechanism. The ratio of elastic zone (A) to the length of disloca­tion scatters (L) decreases by two times with temperature rise from 300 to 1,080K through dislocation motion and relaxation of elastic stresses. Observation for changes

of elastic distortion and dislocation motion after annealing of carbide samples with indentations at various temperatures permits to determine the starting stress as, for dislocation motion and yield stress ay [20]. The as for ZrC and other refractory car­bides is high in the low temperature range of 0.15-0.3 Tm as in covalent crystals (Ge, Si) with high Peierls stress (intrinsic lattice resistance to dislocation motion). ay is three orders of magnitude higher than as. This suggests the controlling rate of dislo­cation generation and dislocation pinning. The value of as for metals is three orders lower than in ZrC and Ge. At temperatures above 0.4Tm, the deformation mecha­nism changes and diffusion rate of metal and carbon atoms increases markedly. An active nonconservative motion of dislocations causes the relaxation of local stresses and subsequent decrease of yield stress. Elongated dipoles and dislocation loops disappear and generation of dislocation networks begins. The further temperature increase up to 0.6 Tm brings into existence the cell appearance. The formation of

dislocation in carbides and metals during deformation are very similar. However, the temperature levels for initiation of dislocation motion and formation of cell structure are higher for carbides.

The fuel and structural NRE materials prepared predominantly by the methods of powder metallurgy have many structural defects, since the level of their strength is quite uncertain; varying within 15-25 % nears the mean value. The strength level depends on the loading method; but the variation coefficient W is in fact independent of the loading method. Changes in the strength of single-phase and heterophase car­bide materials in passing from tension to bending calculated by the Weibull method are in good agreement with experimental data. This circumstance is taken into account in determining the reliability of mechanically loaded NRE constructions.

Dispersion of strength in an engineering practice most often is appreciated by Weibull distribution function:

where a > an and an is a stress below which the probability of fracture is equal to zero, irrespective of the dimensions of a body; a0 is the normalizing parameter; m is the parameter describing uniformity of a material, i. e., degree of distribution uniformity of defects on the body’s volume. The distribution of stress estimated on the basis of a bend-and-torsion test of hundreds of samples of ZrC and NbC submits to the normal law and is described conveniently by the Weibull function.

7.1 Fracture Criteria of Thermally Loaded Bodies

Temperature stresses inevitably appearing during the NRE operation are one of the main factors that can cause the destruction of ceramic units of HRAs [1]. In some cases, temperature stresses determine the construction features and output parameters of a device being developed and the possibility of using one material or another. Therefore, to estimate the strength, it is necessary to determine the level of these stresses and the degree of their danger for individual elements and the construction as a whole.

As a rule, the strength is calculated by using admissible normal or tangential stresses that are safe for the product strength or the limit number of loading cycles. For a newly created class of machines or apparatuses, it is necessary to determine the properties of the collapse of materials from which their bearing elements are made and to determine certain criteria for the strength and its limiting values. The strength of a metal HRA housing is estimated by special methods for calculating the strength and the radial and longitudinal stability of cylindrical shells known in rocket building. The bearing capacity of ceramic fuel elements with a cross-section of a complicated shape was estimated by using the strength criteria for bodies operating in the inhomogeneous field of thermal stresses developed at the institute Luch.

The results of studies initiated at the institute as early as 1973 showed that the use of methods of thermal action in different combinations and varying the shape and size of a body changed the thermal strength and the type of body collapse [3]. Based on the concepts of force fracture mechanics, a new criterion factor N was introduced [2], which took the stress distribution into account and determined conditions of the total or partial fracture of bodies upon changing this stressed state. The values of the parameter N for different types of thermal loading were calculated numerically.

Upon heating the side surface of a body made of an elastic brittle material, a crack appearing in the central tensile region causes the complete fragmentation of the body when the critical stress intensity K1c and values N > Ncr are reached (Fig. 7.1a). When the side surface of a heated body is cooled (under conditions similar to the

A. Lanin, Nuclear Rocket Engine Reactor, Springer Series in Materials Science 170, 89

DOI: 10.1007/978-3-642-32430-7_7, © Springer-Verlag Berlin Heidelberg 2013

operation conditions of HREs according to the scheme presented in Fig.7.1b) and the inhomogeneity parameter is N < Ncr, the nonequilibrium propagation of a crack started in the tensile region at the critical value K1c changes to its equilibrium increase. The total fragmentation of the body becomes possible when stresses (after their substantial redistribution) are 8-10 times higher than the start stress of the crack [2]. In this case, the penetration of the crack into the body is 0.55R on average, which agrees with calculations.

The tests of heated HREs of different compositions by the method of nonstationary cooling in water showed that they were partially damaged due to the appearance of surface cracks at stresses exceeding the tensile stress at corresponding to the heat flow qs и 2.5MW m-2 by only 15-20 % [3]. The bearing capacity of the HRE estimated from decay in the strength decreases almost three times after the appearance of surface cracks and remains virtually unchanged after repeated cyclic loadings. We note that the number of cracks per unit surface increases when qs increases, while their penetration depth in the body and hence the bending strength do not change (Fig. 7.2). Tests of HREs by passing an electric current through them and blowing off their surface by a gas flow at surface temperatures (1,500-1,900 K) higher than upon cooling in water confirm that ZrC+UC HREs were damaged for qs = 2.5-3.0MWm-2 (Fig. 7.3). Doping a carbide matrix with carbon inclusions almost doubles the damage threshold (up to qs 5MWm-2). The complete fragmentation

of HREs made of ZrC + UC and ZrC + NbC + UC occurs at mean values qs = 10 — 12MWm-2.

The estimation of vibration strength of HRA is made at room temperature without radiation. More than 15th overloading during 50h in a frequency kilohertz range, and singular blows at the case in three directions did not cause any change of HRA state [4].

The working capacity of fuel elements (FE) and HGA in reactor conditions, mod­eling modes of reactors’ operation of the nuclear propulsion reactor (NPR), and nuclear propulsion energy reactor (NPER) on a propulsion mode (PM) with two various power levels implemented in reactors IVG.1, IR-100 and on an energy mode (EM) in rector RA [6, 7] (Fig. 7.3).

NP tests were carried out in two HGA modifications cooled by hydrogen techno­logical canals 300 (TC-300) and experimental technological canals (ETC) of reactor IVG-1, FE tests implemented in suppressed and filled with helium canals TC-100 of the reactor IR-100 and in filled helium ampoules of reactor RA. In total, 152 techno­logical canals of various type have been tested on PM. Reactor tests of the irradiated FE of various heating sections (HS) of HGA and FE of the resource ampoules were examined. In total, 110 various types of canals after reactor tests were investigated.

Fuel elements of nuclear rocket engine (NRE) were tested in IVG-1 research reactor in the modes that simulate operating conditions of NRE reactor. These fuel elements were used in fuel assemblies (FA), containing several heating sections (HS) 100 mm long inserted into the flowing gas-cooled technological channels of NRE reactor core.

During the working out of NRE fuel elements in IVG-1 reactor, 30 technological channels with 8- or 6-cell fuel assemblies were installed. Each HS in the 8-cell assemblies contains 379 fuel elements, and each HS in the 6-cell assemblies—151 fuel elements (beam diameters of fuel elements in the 8-cell and 6-cell assemblies were respectively 47.0 and 29.7 mm). In the 8-cell assemblies the first five HS contain fuel elements of (Zr, U)C and the last three are fuel elements of (Zr, Nb, U)C; in 6-cell assemblies, the first four HS contain fuel elements of (Zr, U)C+C, and the last two are fuel element of (Zr, Nb, U)C.

During the ground test of NRE fuel elements each IVG-1 reactor start-up (of about 5-6 min duration) corresponded to one of the reactor activation of space NRE. In the trial tests of the IVG-1 reactor (first core) only three start-ups were performed—power start-up (PS) and two working start-ups (WS). And in the life cycle tests of the second core the regulated number of the start-ups was made (that is, 1 PS and 11 WS). At that the modes of WS in the life cycle tests corresponded well to operating conditions of the reactor activations of space NRE. In particular, the hydrogen temperature at FA outlet reached 3,100K. Hydrogen pressure at the inlet and outlet of FA was ~10 and ~5 MPa, respectively, the maximum temperature drop along the radial cross­section of fuel element reached 250 K.

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

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

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

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).

The strength of ceramics is dispersed in brittle condition in considerably greater degree than the strength of metals. This feature is defined, first of all, by variation of both superficial and volumetric deficiency and lack of stress relaxation on stress concentrator’s influence on structural strength parameters.

The strength of ceramics in brittle condition is defined predominantly by the surface and volume defects [3, 14] and an alloying with formation of solid solu­tion has practically no influence on the strength. Severe surface relief on sintered ceramic samples or microcracks appearing after electromachining or diamond cut­ting is responsible for the low strength level [14]. Elimination of these surface defects by mechanical grinding and polishing increases the strength by 50-70 % leaving the strength variance practically constant at the expense of volume flaws. The volume defects in the form of large rounded flaws with sizes of 100-400 ц are responsible for 40 % of all the cases of fracture in monophase ceramics. In some cases, the fracture occurs at lower stress at the cost of zone peculiarities appearing during inhomoge­neous forming and subsequent inhomogeneous shrinkage during sintering.

Severe surface relief on sintered ceramic samples or microcracks appearing after electromachining or diamond cutting is responsible for the low strength level [14].

Elimination of these surface defects by mechanical grinding and polishing increases the strength by 50-70 % leaving the strength variance practically constant at the expense of volume flaws (Fig. 4.13).

Introduction of more dangerous stress concentrators on a surface, than in volume of the sample, by an indentation, reduces the average strength level but decreases dispersion sharply.

The dangerous defects in the form of cracks should be estimated by an X-ray method based on a measurement of a broadening of X-ray lines at the application of load to the sample, placed in a special prefix to a diffractometer [23] (Fig.4.14).

The volume defects in the form of large rounded flaws with sizes of 100-400 ^m are responsible for 40 % of all the cases of fracture in monophase ceramics. In some cases, the fracture occurs at lower stress at the cost of zone peculiarities appearing

during inhomogeneous forming and subsequent inhomogeneous shrinkage during sintering (Fig. 4.15).

These zones in the form of friable porous layers lower the strength by 30-50 %. Maximum of strength and Tb—d is increased by 300-400K in comparison with ceramics having the same porosity level but without these zones [12, 14]. In most cases, sources of failure of ceramic materials are small cracks faintly visible at the magnification of 1,000 with a radius many orders of magnitude lower than for large pores; so, the former are able to cause more dangerous stress concentration.

The availability of pores decreasing the body’s cross-section clearly reduces the strength. The pore size increased from 3 to 90 ^ and emergence of elliptical pore form at a constant porosity level led to strength reduction of ZrC under bending from 300 to 170MPa with constant porosity level of 5-7 %. The exact influence of porosity on the strength is difficult to determine as in most cases, simultaneously with porosity, other structural parameters are varied: grain size, surface and volume defects, and impurity segregation [3].

Structural parameters depend to a great extent on the chosen technological para­meters. The presence of pores in a material naturally reduces the cross-section of a body and its strength, which can be described, for example, by an empirical relation like

a = a0exp(—BP),

where o0 is the strength of a body without pores and B is a coefficient depending on the pore size and configuration [24]. The temperature of brittle-ductile transition of ceramics versus porosity increases. For example, Tb-d of ZrC, NbC at raise of porosity from 7 to 60 % increase Tb-d by 400 K at a bend test and by 600-700 K at compression (Fig.4.16). The maximum strength, owing to suppression of a plastic deformation is decreased and shifted toward to higher temperatures. Similar changes occur at carbide-graphite compositions.

The relationship between strength and grain size is rather complicated due to simultaneous variation with grain size of flaws, segregation additives on the grain boundaries and ratio of grain and boundary volumes in ceramics [2, 3, 14] (Fig.4.16).

The primary recrystallization (annealing of material after preliminary deforma­tion) makes it possible to vary the grain size of ZrC in the wide range from 5 up to 2,500 ^ without altering the boundary conditions. In this case, the variation of grain size retains carbide strength at 280 K constant while a decrease of strength of ZrC samples is observed after accumulative recrystallization (Fig. 4.17). The latter is more common for ceramics, as mentioned in references. The yield stress o0.2 of ceramics follows Hall-Petch law in the temperature range 0.6Tm <T> Tb-d [25].

00.2 = 00 + kydg-1/2

where o0 is the Peierls stress and ky is the coefficient of deformation resistance through grain boundaries. The relationship is altered radically at temperatures above 0.6 Tm. The yield stress o0.2 decreases owing to the grain boundary sliding and rotation of crystals.

Microstructural defects such as pores, grain boundaries, and microcracks located between grains or phase components influence the resistance to cracks for materials prepared by powder metallurgy methods. The value of K1c for carbide single-phase materials does not exceed 3MPam1/2. As a rule, K1c no monotonically changes with temperature. First, K1c decreases or remains unchanged with increasing the

temperature and begins to increase only when plasticity develops. The effective surface energy depends on the environment in a complicated way.

For exclusion of the most defective work pieces and the decrease of strength dispersion, it is expedient to carry out rejection by a preliminary mechanical loading [6]. For achievement of a positive effect of the rejection, it is necessary for each product to install the mode and level of a loading on the basis of the statistical data about strength distribution taking into account a possible damageability and economic feasibility (Table4.5).

Operating conditions of fuel elements in NRE reactor (and test conditions of NRE fuel elements in IVG-1 reactor) are associated with essential presence of thermal stresses in them, because all the fuel elements heat from the internal energy is selected by the hydrogen flow pumped through the fuel assembly. The most intense heat removal from the surface of the fuel element (creating a radial temperature drop AT over its cross-section) takes place in the inlet HS, operating in the temperature range of brittle state that means the first and second HS in the 6-cell FA or the first, second, and third HS in the 8-cell FA.

In the course of the reactor steady-state operation, the temperature drop in each fuel element increases from zero to maximum and remains at that value during the

Fig. 7.4 The radial distrib­ution of temperature T, the axial thermal stresses Ot and axial residual radiation stress Or resulting in NRE fuel ele­ments at various stages of its operation in flowing techno­logical channels

start-up (the parabolic form of the temperature distribution in cross-section of fuel element is shown in Fig.7.4). Temperature drop value is determined by the heat flux qs from the surface of a fuel element, thermal conductivity X of the fuel element material and the radius r. This temperature drop AT creates dangerous thermal macro stresses ot with tensile component at the fuel element surface (see the diagram of axial ot in Fig.7.4). The values of the axial and tangential tensile stresses ot can be found [8] from the formula:

ot = aEAT/[2(1 — v)]= a Eqsr/[4X(1 — v)], (7.1)

where a—coefficient of thermal expansion; E—Young modulus; v—Poisson’s con­stant.

Thermal stresses ot pose a real danger for fuel elements operating only at tem­peratures up to 1,600K. Because in these conditions, the brittle refractory carbides will not have relaxation capability of these time elastic macro stresses. If the stress ot exceeds the tensile strength of fuel element material there will be cracks on its surface. Cracks in themselves do not affect the operation of the fuel elements, but their appearance leads to very negative consequences, as in the future failure of fuel rods weakened by cracks by bending, bandage and vibration loads (bending loads occur in the fuel elements in the event of the longitudinal beams twisting in HS, and bandage—under the influence of radiation thickening of fuel elements).

Along with the negative aspect of hazardous thermal stresses presented in the fuel elements there is a positive sign consisting of fuel elements hardening due to the appearance of compressive residual radiation macro stresses on their surfaces or (see diagram of axial or in Fig.7.4). Residual stresses or arise again because of the presence of radial temperature drop AT in the irradiated fuel element, which provides radiation swelling heterogeneity of the fuel element material. Due to different rates

of annealing of resulting radiation defects volume of peripheral (cooler) regions of the fuel element will increase stronger than the internal volume of (hotter) regions. This will ensure the appearance of compressive residual stresses on the fuel ele­ment surface and tensile stresses inside, as the internal regions preventing excessive expansion of the surface regions, will create an elastic compression in them.