Model tests of the component units of HRAs are, along with calculations, the main tool for the experimental substantiation of the construction at the stage of technical project development. The tests were performed with model samples of component units under laboratory conditions. A specific feature of these studies is the separate influence of different action conditions. The complex action was estimated using the superposition principle by alternating separate tests in stages. This course of action is admissible in the case of the block HRA construction and is caused by the complexity of simulating natural operation conditions. The results obtained on high-temperature testing units allowed optimizing the design and technological solutions used in the HRA development.

At the initial period of NRE development, estimating the efficiency of the con­struction units of the NRE core was difficult because of the lack of sufficiently complete information on the physical and mechanical properties of materials, espe­cially at high temperatures. It was necessary to verify the correctness of the design and technological solutions in laboratory studies and prereactor tests. To solve these problems, a structure of the experimental equipment was created at the RIHRE, some instruments of the structure being unique. During the development of methods for testing refractory ceramic materials in a broad temperature range from 300 to 3,000K, a number of difficulties related to the specific properties of these materials were encountered.

The high hardness and high brittleness of carbides, nitrides, and other refractory materials complicating their machining exclude the option of using large samples of a complicated shape because of technological restrictions on their manufacturing. The easy oxidizability of these materials upon heating in air requires performing tests in a vacuum or inert medium and preventing the possible interaction of samples with heater materials.

For these reasons, standard setups for testing metals could not be used and it was necessary to develop or refine a structure of high-temperature setups satisfying these requirements.

A. Lanin, Nuclear Rocket Engine Reactor, Springer Series in Materials Science 170, 21

DOI: 10.1007/978-3-642-32430-7_3, © Springer-Verlag Berlin Heidelberg 2013

The working temperature of materials in different parts of HGAs in NREs in transient and stationary regimes changes in a broad range from 300 to 3,000 K. since it is necessary to know the variations in mechanical and physical properties. For structural and fuel materials of the interstitial phase type, the problem of preventing a brittle collapse is of primary importance. Such collapse is caused by the high brittleness of materials of this class in a broad range of temperatures, loading methods and

rates, along with a relatively low tensile strength (compared, for example, with metal materials). Refractory carbide compounds with a high level of resistance to the motion of dislocations are characterized by no monotonic dependence of the strength [14]. First, the strength is virtually unchanged with increasing the temperature and begins to increase only when microplasticity develops, achieving a maximum (Fig. 4.18) at the brittle-ductile transition temperature Tb-d, due to the thermally activated relaxation of local peak stresses near concentrators-structural defects contained in the material. As the temperature was further increased (T > Tb-d), microplasticity developed in ceramics and their strength monotonically decreased, as in metals. The presence of nonmetal (O2, N2, C, Si) or metal impurities (Fe, C, Ni) due to the formation of phases and low-melting eutectics can change the transcrystalline type of collapse to the intercrystalline one and also change the temperature dependence of the strength.

The brittle-ductile transition temperature in ceramics, unlike Tb-d in metals, strongly changes (by a few hundred degrees) upon changing the type of the stressed state and the deformation rate. As the deformation rate of ZrC rise under tension from 10-3 to 10-1 s-1, the temperature Tb-d increases from 0.6 to 0.8Tm. In the case of shock loadings with rates exceeding the relaxation rates of micro — and macrostresses, the strength remains almost at the level of its value at room temperature. A similar behavior is also observed for oxide materials and weakly plastic Si and Ge metals with strongly manifested covalent bonds (Fig.4.19).

Transition temperature Tb-d for complex solid ZrC + UC, ZrC + NbC + UC solutions is higher than for the single-phase ZrC.

Long-term strength with increase of loading time drops. Time of fracture т and stationary creep rate є’ are connected with applied stress о. Laws of creep kinetics inherent to metals are observed and for ceramic materials [22, 26]. In a general view dependence of creep rate є’ on applied stress о can be written down:

t = B ■ am ■ exp(U/RT) є’ = A ■ an ■ exp(-Q/RT)

where A, B, n, m—constants, Q and U—activation energy of creeps and long frac­ture processes. For carbide materials tested in a temperature band 2,500-3,000 K, activation energy Q and U and exponent’s n and m of stresses practically coincide.

Creep rate depends from a mostly under the linear law in an interval of small stresses 10-20MPa at high temperatures 0.6Tm. According to this, we usually dis­tinguish threshold (i. e. developed at stresses surpassing some stress a > at) and not threshold creep; the last can be observed at any as much as small a. No threshold creep is carried out by the directed diffusion of atoms and proceeds on vacancy-diffusion mechanism of Nabarro-Herring. Creep rate inversely proportional to a square size of grain and in a strong degree depends on stochiometric of the compound (Fig. 4.20a).

At stresses a > at alongside with diffusive processes, conservative movement of dislocations is initiated and dependence є’ from becomes power mode є’ ~ a(3-4) (Fig.4.20a) and the intergranular sliding at the final stage of creep generate microp­orosity on the grain boundaries that causes appreciable decrease of density. Generally, creep deformation is caused by passing not one but several processes (diffusive flu­idity. sliding on grain borders, dislocation moving), the contribution of each process depends on conditions of carrying out of experiences (temperature, stress, duration) and structural characteristics of the object.

The instantaneous and unsteady stages of creep are absent in most cases at stresses greater at, but smaller yield point a0,2. Accumulation of a plastic strain basically occurs at stages of the stationary at and accelerated creep, and duration of a final site leaves approximately 1/3 from time before fracture. The general accumulated plastic strain during creep

where є’—creep rate on the stationary stage; єш—deformation on the accelerated stage; t—time.

The structure of ceramics changes with increase of a creep strain, agglomeration of microspores in the form of chains on borders occur in grains, perpendicular to action of loading. Structural changes result in to change of density and physical properties. Increase of porosity, for example, in beryllium from 1.4 up to 12% raises є’ in 2-4 times. Generally, porosity dependence of creep can be expressed by empirical expression.

є’~ (1 — p2/3)-1 or є’~ (1 — p)-3’ where p is porosity of a sample.

The fatigue behaviors of zirconium carbide [14] under cyclic loading have demon­strated that fatigue cracks in ceramics can grow at lower stress than static strength. The cyclic tests were carried out under compressive loading at room temperature in ambient air. Subcritical crack velocity is described by:

dl/dN = a(Ki /Kic)n

where Ki is a current stress intensity. Kic is the fracture toughness, a and n constants depending on type of loading and the environment, Kic is determined by the developed method under uniaxial compression for plane specimens with a central cut at a certain angle to an axis. The crack starts at Kic and after sporadically increasing length stress intensity drops to a lower Kia value. The subsequent crack growth is possible only at increasing external load, these permits to multiply determination of the fracture toughness on one specimen. The results of cyclic tests show that crack propagates even at Ki < Kia with velocities 1-3-10-3 mm/s. The value dl/N increases to 1 mm/s at Ki > Kic, and n value in the range 75-89. The observed subcritical crack growth under compression at room temperature is connected with accumulated localized microcracking and crushing of fracture-surface roughness at the unloading stage. Internal local stresses play an important role in the fatigue behavior of ceramics. The origin of these local stresses depends on heterogeneity of chemical composition, structure, and anisotropy of sintered ceramic blanks [14]. Cyclic loading of cylindrical ZrC samples at room temperature does not decrease the strength ab under constant bending stresses equal to 0.8-0.9 of the mean, ab for a number of cycles up to 106. The cyclic bending of ZrC samples at temperature above 2,300 K increases the strength because of viscous intergranular flow and stress relaxation of local stresses.

The resistance of ceramics to the propagation of cracks can be strongly influ­enced by the microstructure but optimization of fracture toughness and strength usually involves different choices of microstructure: individual toughening mecha­nisms including phase transformations. Microcracking, twining, ductile reinforce­ment, flber/whisker reinforcement and grain bridging as regarded thoroughly in [3]. The optimum structure of ZrC with small grain, pore size below 1 ^ and medium porosity of 8-10 % has the maximum Kic = 3.3MPa — m1/2 and bending strength near 650MPa [14]. The Kic of most refractory carbides, nitrides, and borides does not exceed 5 MPa-m1/2 [2]. The polycrystals based on two modifications of boron nitride with large internal local stresses have extremely high Kic up to 18MPa — m1/2 [2].

Trial and life cycle tests of NRE fuel elements in flowing gas-cooled channels of IVG-1 research reactor showed that the state of the irradiated fuel elements to a great extent depends on the specific modes of reactor tests [9-11]. Thus, the fuel elements of (Zr, U)C and (Zr, Nb, U)C after the trial tests in the channels with 8-cell fuel assemblies in the modes, significantly reduced compared to the nominal mode for NRE reactor, were discovered in a quite satisfactory condition: surface cracks and fracture of the fuel rods were few events, the longitudinal beams twisting of fuel elements in HS were small, the increase in the coefficient of gas-dynamic resistance of the cooling duct of fuel assemblies was also small. At the same time a significant post-irradiation increase of strength aS and thermal strength of fuel elements, as well as the presence of compressive residual stresses aR on their surfaces was recorded.

Hardening of fuel elements AaS/aS in three inlet (low-temperature) sections (HS-1, HS-2, and HS-3) reached almost 100%, in four outlet (high-temperature) sections it was small (Fig. 7.5).

The nature of fuel elements thermal strength increase was the same as the char­acter of their hardening. Measurements of fuel elements thermal strength by time — dependent thermal loading (destructive temperature drop ATF or the first criterion of the thermal strength was determined by rapid immersion of the pre-heated fuel rods in water) showed the more than twofold increase in thermal strength of fuel elements of (Zr, U)C in the first three inlet HS and the almost complete no change of

the characteristics of fuel elements of (Zr, U)C and (Zr, Nb, U)C in high-temperature HS. The coincidence of the change of the strength and thermal strength of fuel ele­ments on FA length is due to the fact that the thermal strength of brittle materials products is linearly related to their strength.

The results of X-ray measurements of the magnitude and sign of axial residual stresses or in fuel elements of two channels exposed to different uranium burn-up ND (ND burn-up—the number of uranium fission per a volume unit) are shown in Fig.7.6. In the measurements, we used the developed nondestructive method [12] with photo recording of the stresses in a special X-ray camera. The measurements were made on the fuel elements which had no surface cracks. When comparing the data in Figs. 7.6 and 7.5, we can see the satisfactory correlation of the change in hardening levels of fuel elements along the FA length and the levels of compressive residual stresses in them: the maximal hardening Aos/os and maximum stresses or occur in the fuel elements of the first three HS of both channels.

Furthermore, using only the fuel elements from HS-1 sections of six channels. X-ray study of the uranium burn-up effect on the level of occurred stresses or was conducted out. It turned out that ND increase on an order (from ~2.1-1016 to ~2.1-1017 div/cm3) leads to the more than twofold increase of stresses level or. This has led to express concern for the integrity of the fuel rods involved in a longer state life tests at IVG-1 reactor with the second core. Indeed, a further significant increase of ND burn-up the positive effect of increasing compressive stresses at the surfaces of fuel elements may be replaced by the negative: a significant increase of the tensile axial stresses or inside of fuel elements could lead to their damage into fragments at the time of cooling after start-up (at the time of temporary stresses disappearance ot).

The question on the danger of expected excessive growth of stresses or in the fuel elements was resolved only after the completion of their life cycle tests in the second IVG-1 reactor core, when the opportunity to X-ray measurements of stresses or in fuel elements with ND burn-up in a wide range (from ~2-1016 to ~2-1018 div/cm3). To do this, in addition to data on the fuel elements of (Zr, U)C tested in the six channels of IVG-1 reactor first core, X-ray diffractometry parameter measurements

of the magnitudes and signs of axial residual stresses or on the surfaces of fuel elements of (Zr, U)C+C tested in low-temperature HS of 11 channels of IVG-1 reactor second core were performed. The measurements were carried out on the selected integer fuel elements containing no surface cracks.

The results of measurements or (see Fig.7.7 in which data for the first core fuel elements are marked with square dots) indicated that the continuous growth of the radiation dose up to uranium burn-up about 2-1018 div/cm3 was not accompanied by continuous growth of compressive stresses or on the surface of a fuel element (and tensile or inside) to excessively high values. The increase of stresses or (fuel element material swelling heterogeneity increase), as seen from the figure was stopped after burning of about 3-1017 div/cm3 (indicated by dotted line) before reaching 200MPa, that eliminates the risk of failure of fuel elements from the tensile residual stresses operating in their internal regions.

In contrast to the first core of IVG-1 reactor, the second core fuel elements tested in modes close to the nominal operation of NRE reactor, there was a significant degradation of the initial state, especially noticeable in low-temperature HS-1 and HS-2. In particular, the fuel elements in these sections had numerous cracks and damage, and the longitudinal beams twisting in HS and the numbers of failures of fuel elements were increased with increasing the number of reactor start-ups. For example, after six start-ups the angles of beams twisting of fuel elements reached about ~20°, and the number of broken fuel elements (Fig.7.8) was approximately 80%.

Destroyed fuel elements in HS-1 and HS-2 were small (3-15mm in length) and slightly mixed fragments of rods that significantly—hundreds of percent—increased the coefficient of gas-dynamic resistance at the entrance of FA cooling duct. The fuel elements in the remaining four high-temperature HS remained intact, or were damaged only to large (30 mm length) fragments without mutual displacement, their strength decreased with increasing temperature and duration of exposure due to the degradation of the surface material by erosion and hydration.

It is obvious that the original cause of the negative situation in the inlet HS is the lack of thermal strength of carbide-graphite fuel elements at the temperature of the brittle state. Indeed, the level of thermal strength of these fuel elements is such

Fig. 7.8 View of fuel elements fracture from the second heating section (HS) (integrated flux of fast neu­trons of order 1019sm-2)

that the cracks appear in them even at heat loads qS higher ~5 MW/m2 (see upper dangerous region in Fig.7.9). That is why the fuel elements of the first two sections of 6-cell FA were damaged by cracks and destroyed during tests in IVG-1 reactor second core, and the fuel elements in the first three sections of 8-cell assemblies remained intact during the tests in the first core.

The cracks arising in FE under the influence of thermoelastic stress lead to a final fracture in the first and second HS by a tightness of ten-order microns on a series of the peripheral mainly damaged FE. Generated FE fragments can drop out and increase hydraulic resistance of HS, in a case if they have moving freedom in any direction. It depends on a concrete locating of fragments and also on angular orientation surrounding the yielded FE fragments and forces, capable to move them (Fig.7.10).

Fuel elements in the last HS undergo plasticity at elevated temperatures, forming sometimes through axial channels (Fig. 7.11). The decline of the heat removal, due to change of gas flow in the channel, leads to temperature increase of fuel elements that in the presence of compression loading promotes emersion of plastic deformation.

FE of the last high-temperature HS can also be twisted around axis and lose the stability in the presence of axial compression load under the influence of axial forces of the downthrusted spring and power loads, arising at gas passing in HS (Fig.7.12).

Consequently, for the elimination of a failure of fuel elements in the first two sections of 6-cell assemblies and prevent processes of grinding and mixing of the fragments it is necessary, as seen in Fig.7.9, to increase the initial thermal strength of the fuel elements of (U, Zr)C+C approximately double. As the level of thermal strength of brittle carbide material can be elevated by increasing its strength, in [10] the possible ways of pre-hardening of fuel elements by thermal, radiation, and combination methods were analyzed. Among the possible radiation methods, the way of fuel elements strengthening by residual radiation stresses or, in which the fuel elements of 6-cell FA were previously irradiated to ND burn-up about 3-1017 div/sm3 in a mode of decreased power (at average heat load on FA length qS about 4MW/m2), was pointed to. Pre-irradiation is carried out in the shape of the first start-up (FS) of NRE space facility located in orbit. Double hardening of carbide-graphite fuel elements of low-temperature HS achieved during FS must ensure the absence of cracks and breakage in the subsequent reactor activations.

A comparison of the change of fuel elements strength after reactor tests in only a few (n) working start-ups (nWS) and after the tests in the n working start-ups with prior power start-up (PS+nWS) [10, 11] can serve as an experimental substantiation of the proposed method.

Such positive effect of PS start-up was recorded among the fuel elements tested in the IVG-1 reactor fourth core (see Fig.7.13): hardening of fuel elements in HS-1 and HS-2 after the tests in PS + 2WS and 2WS amounted to ~35 and ~20 %, corre­spondingly.

Special setups were developed for measuring the strength and plasticity of ceramic materials during bending, compression, tension, and creep and the endurance limit at high temperatures up to 3,000K in inert gasses or in a vacuum (Fig. 3.1).

When using the dynamometer, the relative measurement error of stresses rang­ing 0.4-1,000MN does not exceed 1 %. The measurement error of sample strain, estimated as support displacement difference, was 5%. This system was used for Russia’s first strength test for graphite and ZrC at 3,000 K in 1960. The PRV-203 system built by the Research institute of test machines (RITM, Moscow), originally developed for testing of large samples heated by tungsten heater up to 2,500 K in vac­uum (see Fig. 3.2), was subsequently reworked (in particular, the clamping fixture and dynamometer were redesigned) for bending test of smaller samples (2 x 2 x 25 mm) with stress measurement error of 2 % at the lower load limit of 0.2 MN. Sample strain (with 5 % measurement error) was estimated by support displacement as registered on the device diagram tape at a scale of 1:100.

In 1973, ordered by “Luch” the RITM has developed the electron beam machine EBM-402 for mechanical testing in vacuum at temperatures up to 3,300K. Sample surfaces were heated by scanning electron beams of two electron guns (see Fig. 3.2a). Samples were subjected to the load of 2-2,000 kgf with adjustable loading speed of 3 x 10-5 to 0.4 s-1. Sample strain was measured by extensometer connected to

variable-induction sensor. Sample temperature was measured by the optical pyrom­eter, whereas the local temperature pulses in multiple points of the sample were registered by the photovoltaic device. The measurement error of the true sample temperature at 3,000K was ±86K.

The thermal stress resistance (TSR) is the most important characteristic of the effi­ciency of HRA construction elements [6, 14, 27, 28].

A thermal elastic stress can be easily determined at known temperature patterns from combined equations of stress and strains with consideration of thermal defor­mation. Dependences between shear stresses and strains, and also the static and geometrical equations are accepted without modifications on theory of elasticity [24, 29].

The thermal stresses a can be defined by following relationship:

a = E ■ a — ATm ■ Kf/(1 — ^),

where ATm = (Tm — T) is a calculated mean integrated temperature difference ATm = (Tm -1) at the moment of fracture (where Tm is a mean integrated tempera­ture of a body cross-section and T is surface or center point temperature, assuming that properties of a material do not vary and material is elastic), a is a coefficient of thermal expansion, E is a modulus elasticity, ^—is Poisson’s ratio, Kf is form factor. Derived from the relationship at value Kf = 1 complex a(1 — ^)/E — a = ATm is accepted to

Fig. 4.21 TSR for ZrC disk (30 mm in diameter with thickness 2-3 mm) as function of temperature and heating rate. (1) 10K/s, (2) 50K/s,

(3) 100 K/s, (4) 200K/s

name the first criteria of thermal stress resistance R = at (1 — \,)/at E measured in degrees. The complex R’ = XR is a second criteria, which reflect ability of materials to maintain a thermal flow at stationary or nonstationary thermal influences.

The higher the heat conductivity X of a material is the greater the thermal flux that can be withstood by the body without its collapse. It should be noted that the choice of materials for heat insulators in constructions based on the value of R’ is meaningless. In this case, it is necessary to use the value of R for estimates because, given two materials with identical heat-insulating properties, the better material is the one capable of withstanding a greater temperature drop.

The TSR of most refractory ceramics lie in the range ofR = 30 —100 K (Table 4.6). The criterion R’ can change in more wide limits 100-1500 W/m. The thermal stress ath (causing appearance of first cracks) is slightly exceeds the tensile strength and ath/ab, lies in the range of 0.45-0.57 of bending strength value.

The knowledge of TSR versus temperature is necessary for the appraisals stability of ceramic articles under working conditions. The temperature dependence of TSR has three characteristic intervals [6]. The TSR of ceramics is invariable in the elastic — brittle temperature range. A relaxation of thermal stresses increases the TSR with temperature rise in the second interval. The tests performed on ZrC disks by induction heating method [6] show this possibility (Fig. 4.21).

The R value of ZrC remains constant with temperature increase up to 1,700 K regardless of heating rate. The TSR criteria R and R’ of a porous ZrC considerably change when the materials porosity varies (Fig. 4.22). Pores reducing the cross­section area of a body cause a drop in strength a and elasticity modulus E. The a and E dependencies on porosity P may be described by empirical relations of this sort:

a = a0exp(—BP), E = E0 ■ exp(—BP).

The invariability of TSR in the range from 500 to 1,700 K is conditioned exclu­sively due to the constancy of strength in this range, since the E decay and a rise

with temperature compensate for each other and the product aE is kept practically constant. The lower is the velocity of thermal loading above 1,700 K the greater is the reduction of the local stresses and the greater is the up growth of TSR. The thermal fracture in the third temperature range cannot be observed when the thermal stresses relax more rapidly than their formation. The ZrC samples do not fracture even at 1,900 K at low heating velocities <10K/s and much less at the onset of the macroductility at 2,000-2,200 K.

Distribution of R values for ZrC, NbC measured on a representative number of samples (70-90) is characterized by exponential Weibull relation. The coefficient of variation W = S/Rm (S is deviation of TSR, Rm is average value) agrees closely with appropriate significance of the W value obtained under mechanical test of strength. The cyclic thermal loading does not decrease the carrying capacity of samples in elastic-brittle temperature range. TSR of the ceramic bodies can be seriously changed under combined influence of the thermal and mechanical loads [28].

TSR of the ceramic materials can be essentially raised by introduction of hard inclusions in a matrix. The introduction of second phases as carbon inclusions with the sizes varying from 0.5 up to 50 |x into ZrC and NbC matrixes inevitably reduces the strength but is capable to increase the fracture toughness by a factor 1.5-2 with an optimum choice of concentration and kind of carbon inclusions. With growth of the carbon content, the TSR either increased in case of use of carbon black or decreased for graphite [6]. TSR of compositions with carbon black can be raised 2.5-3 times (Fig.4.23) due to an increase of relation a /E [6].

The volume content of the inclusions, their geometry and dimensions have also a notable effect on the TSR as shown by ZrC-compositions with diamond particles [17].

Use of synthetic diamond powders with a narrow spectrum of grains get to obtain the compositions (after sintering) with guaranteed dimensions of the carbon. The inclusions are distributed in the matrix uniformly due to a good mixing of the ZrC and diamond powders. In addition, the diamond volume change during the structural transformation has an active effect on the process of the composition compaction when sintering. With introducing small carbon particles (0.6-6 |x) up to 5 mass % of the strength and the fracture toughness are not practically changed but the modulus of elasticity is decreased considerably. Introduction of 50 |x particles results in simul­taneous drop of a/E and Kic. Inadequate change of the strength and the modulus of

elasticity when increasing the volume content of the diamond particles, results in the a /E-value growth. Therefore, the TSR of the compositions with the diamond parti­cles 0.6-6 ц in size is increased to a greater extent. Apparently, some increase of the fracture toughness does not influence the TSR considerably. The TSR of composi­tions with carbon inclusions is left invariable in temperature range 400-1,900 K and the advent of the first evidence of macroductility happens at more higher temperature up to 2,400 K.

Reactor tests of the HIP, in general, confirmed the results of laboratory tests on the electron arc installation (EAI). The traces of interacting working medium with the surfaces of the TIP from low density graphite (LDG) were detected. Intensity of

hydrogen interaction increased with growth of temperature along gas traffic route. The fourth heating section (HS) was observed sometimes through holes in HIP. Material ablation in these HIPs made up an order of 25-50weight%. The most defective was outer range of TIP from high density graphite (HDG).

Casings from ZrC with Pq flaking forming thermal insulation in high temperature part of HRA, in general, kept the geometry though the casings have some cracks, and at dismantling they scattered on separate fragments. The carbide casings also undergo some changes. There were separate cracks and grids of cracks. The overhead part of the HIP near to the supporting exhaust knot is underwent changes also (Fig. 7.14).

Thus, the basic defects of HIP were the cracks and an area of material flushing. The last is connected, both with chemical interacting and with an erosive material removal. The most dangerous from the point of working capacity is the ablation of HIP under contact of hot working medium with bared metal casing. The structure imperfection of HIP is increased with operating time. There is great scattering between alternative technological versions of HIP preparation. More optimum materials’ composition and reliable design of HIP will be chosen in the subsequent time of NRE development.

The thermal properties in different temperature ranges were measured by different methods using different setups. Comparative estimates of the thermal diffusivity a, the heat conduction X, the heat capacity C, and the electric resistance p were made by fast transient methods with a typical accuracy of ±10% in the temperature range from 300 to 1,000 K. The heat conduction of carbide materials at high temperatures (up to 2,800K) was measured with an accuracy of ±15 % by the axial method by passing current through a rod sample or a fuel element. Simultaneously, the electric resistance was easily measured.

Special technique and equipment were developed for thermal conductivity charac­terization of the multilayer heat insulated packets (HIPs). The radial heat flux method based on indirect internally applied heating (cylinder technique, see Fig. 3.3) yielded reliable data on heat conductivity Xeff and thermal resistance (Rn) of the HIPs by the radial thermal stream method with indirect heating.

The bearing capacity of thermally loaded constructions was estimated by methods with parameters closest to the operating conditions of devices. The thermal loading of an HRE was simulated by cooling the heated HRE in water, while the thermal loading of cases in the HIP was simulated by heating in melted tin [1]. The thermal strength of materials at temperatures exceeding 0.3-0.5Tm (where Tm is the melting point), when micro — and macroplasticity can appear, was measured by electron, plasma, and induction heating methods, which require sophisticated equipment. These methods

Fig. 3.3 The circuit design of assemblage for measurement of heat conductivity TIP.

1 heater, 2, 3, 4 casings: accordingly, internal, testing and outdoor, 5 pyrometric channels, 6, 8 apertures for potential and potentiometer sounds, 7 a collet system and a crimped lock, 9 the pyrometric channel

were used to vary and control the level and type of temperature fields and the rate and cycle of their variation, thereby permitting quantitative estimates of the ther­mal strength of materials under conditions close to operating conditions at high temperatures.

The interaction of the HRA units with hydrogen at working temperatures and changes in the composition, structure, and physical-chemical and mechanical prop­erties were investigated using electrical-thermal setups in which the required tem­peratures could be produced, up to the melting points of materials directly in HREs made of Zr-UC + C and UC-ZrC-NbC. Hydrogen or hydrogen-methane mixture flows passing through working chambers provided the gas rate near the sample sur­face approximately from 0.1 to 40ms-1 at pressures up to 10MPa. The hydraulic resistance of the HRA channel in the construction was optimized on hydraulic mounts (Fig. 3.4).

A unit for thermal heat resistance testing by radiation heating under nonstationary conditions includes, as a rule, a hot (a) and a cold chamber (b, c) (Fig. 3.5). The specimen in the hot chamber is heated in an inert medium or in a vacuum to 2,000 °C by means of heater 1, which is a hollow graphite cylinder placed inside an inductor. The heater is shielded from the surroundings by water-cooled double quartz tubes. In

Fig. 3.4 Circuit design of the working chamber of installation B-2. 1 the tested sample (Fuel element), 2 a quartz tube in diameter 10mm, 3 bearing pipe, 4, 5 current leads with an air-floating head, 6-9 clutches, the 10 steel case

Fig. 3.5 Circuit design of the combined installation RAD for definition of ther­mal strength resistance by a heating (cooling) method of the sample by radiation or cooling in liquid medium [1]. 1 Heater, 2 specimen, 3, 5 shield, 4 piston, 6 sound duct

order to create a uniform temperature field along the length of the graphite heater, the heater’s edges were protected by molybdenum or graphite shields. The temperature of the heater is monitored by an optical pyrometer. After the heater has reached the desired temperature, specimen 2 (protected on the side face by shield 3) is transferred pneumatically from the cold chamber to the hot chamber as piston 4 is driven by pressure applied from beneath. After the specimen has been positioned in the middle of the heater, shield 5 is released and allowed to drop by gravity, and the side face of the specimen is exposed to the heater. The instant at which the specimen fails is recorded by a piezoelectric sensor connected to the specimen through a sound
duct 6; the signal from the sensor is transmitted through an amplifier to a recording oscillograph which is switched on automatically after shield release.

The transfer of heat from the heater at temperature Th to the specimen’s surface at temperature Ts can be treated in terms of constant heat flow q = const with an accuracy better than 5 % if the surface temperature is not in excess of 0.475 Th [1]; this makes determination of the thermal stress resistance a much simpler procedure.

The maximum relative error in measuring the temperature does not exceed ±0.7 %; this error is associated with the thermocouple calibration error (±0.3 %), measuring device error (±0.05 %), and delay effect error (±0.3 %) which is caused by the heat resistance at the thermocouple—specimen contact through an In-Ga eutectic at temperature change rates of <100K/s. As the heat flow for the system “heater — specimen” is known, the integral mean temperature drop to failure is determined by merely measuring the heating time to failure rf.

The total relative error in the thermal stress resistance composed of the time — to-failure error, heat flow error, temperature extrapolation error, and thermal para­meter errors does not exceed 10% with a probability of 0.68. As a rule, failure is manifested by complete disintegration of the specimen.

It should be noted that the cylinders and disks used as test specimens are more convenient than hollow cylinders or rings. However, in both cases heat resistance of the material in terms of mean integrated temperature ATm does not exceed 100 °C, and the temperature at failure can be varied only within 300-700 °C.

Test process consists in a definition of the temperature pattern at the body fracture moment and a calculation of fracture mean integrated temperature drop in a body ATm = Tm — T, where Tm—mean integrated temperature on cross-section of a body and T is a point temperature (on a surface or on a body center), where thermal stress are defined

a = aEATm/(1 — i), (3.1)

where E—modulus of elasticity, a—a coefficient of linear expansion, ^—a Poisson’s ratio.

The complex (1 — i)/aE = ATm is accepted to name as the first criterion of thermal strength resistance R [1].

High-temperature thermal strength resistance under electron beam heating was measured on UTU-1 system (see Fig.3.6) developed at RITM, Moscow. The said technique involves both sophisticated equipment and an elaborate measurement procedure, but its capabilities if high-speed thermal loading of the sample (up to 1,000K/s) in a wide temperature range up to 2,800K [2] led to effective simulation of the operating conditions of some HGA blocs in NREs. The cylindrical samples in the UTU-1 system were heated by four electron beam guns placed in a horizontal plane at 90° in a sealed chamber under vacuum (min. 0.01 Pa). Thermal strength behavior at various temperatures was measured by slow heating of the sample to the preset temperature, with subsequent high-speed heating ramp up to the sample fracture. The fracture criterion used was the abrupt slope change of the temperature curve as registered on the oscilloscope tape. Temperature stress calculated using numerical methods based on the first boundary conditions, with surface temperature

being a given function of time. The general measurement error of the thermal strength resistance was estimated at ±15 %.

Researchers of RIPRA “Luch” have invested considerable efforts into the develop­ment of various alternative manufacturing techniques for processing heat insulating materials (HIP) of the nuclear fuel assemblies (NFA) of Nuclear Rocket Engines (NRE), shows two alternative packaging arrangements of heat insulation packages (HIP) for NFA.

Five-layer HIPs based on three ZrC + C casings Fig.2.9b are designed for the thermal flow values of 2 x 106 W-m-2 [30]. The HIP protects the NFA housing from the heat impact of the working medium. This design features a multilayer sectional package structure that minimizes the possibility of cracks penetrating into the housing and offers the possibility to vary HIP material composition over its length and across its thickness.

The pyrolytic graphite-based outer casings of the HIP, along with good thermal insulation, provide for a ‘soft’ contact with the housing, thereby facilitating the structure assembling process and minimizing abrasion of the metal case by thermal insulation. The inner casings are thin-walled carbide-graphite cylinders. For the low — temperature section these casings are made of zirconium carbide + graphite (ZCG) and for the hot section the niobium carbide + graphite (NCG) casings are used. These casings serve as a supporting frame for the HRA preventing the penetration of fragments of heat insulation elements into the hot sections’ duct. The casings ensure the assembly of HSs and their mounting into a heat-insulating package and reduce the effects of erosion and aggressive impact of the working medium on the heat insulation. Casings made of low-density pyrographite and porous ZCG and NCG and typically placed between metal housings. Casings made of low-density pyro-graphite are located in the low-temperature section (T = 1,500 — 2,000 K). For higher temperatures, the first designs contained casings based on so-called “laminate” consisting of carbide layers in a graphite matrix. The next-generation designs had these replaced by casings made of porous Zr and Nb carbides.

The HIP internal casings are made of ZCG or NCG containing 5 wt% of the free carbon. Use of ZCG is limited by the eutectic melt temperature of compositions. The NCG layer serves as a rigid carcass preventing the possible penetration of HIP frag­ments into the hot section duct. The low-density pyrolytic graphite layers (deposited on the surface of one or two casings) are used in the zones characterized by the max­imum power flux. The length of this zone is limited by the temperature of reaction with the environment (hydrogen).

Preforms of ZCG-based on laminate pyrolytic compositions (ZrC + Pq) were processed by depositing carbon and zirconium particles from the vapor-phase mix­ture of ZrCl4 CH4 + H2 + Ar on the surface of a thin-walled graphite substrate at 1,700-1,800K [31].

Pyrolytic graphite (PGV) is a polycrystalline form of graphite deposited from a vapor-phase carbon source at pressures of 25-150 mm Hg on a heated graphite substrate [32, 33]. The PGV casing is used in the relatively cold zone of HIP. PGV casing provides reliable heat-shielding of the duct housing and prevents metal hous­ing exposure to hydrogen. PGV has a highly desirable feature for the given operating conditions: its thermal conductivity slightly decreases with temperature rise.

Pyrographite properties depend primarily on the substrate temperature, pressure and speed of the flowing vapor and on the composition of a vapor mix. By rising the deposition temperature from 1,800 to 2,500K it is possible to obtain pyro-graphites with the density varying from 1.2 to 2.25g/cm3.

Such density variations lead to significant changes in the physical properties of the pyrographite; thus the density is considered as the basic parameter influencing the main properties, including thermal conductivity. Distinctive feature of pyrographites is their significant anisotropy of properties. In particular, perpendicularly to the depo­sition surface (c axis) thermal conductivity maybe 50-100 times lower than that along deposition surface (a axis); depending on density; the thermal conductivity may vary from 1.5 to 7 W/m-K [33]. Due to the lower thermal conductivity along c-axis pyro — graphites may effectively be used in heat shields for various power plants with high operation temperatures.

The results of analysis of the pyrographites thermal conductivity values show considerable scatter both in absolute values, and in their change with temperature along c axis [31-33]. This further complicates thermal calculations and the prediction of performance of specific thermally stressed parts of HGA. In order to obtain the

Fig. 4.25 Temperature dependence of heat con­ductivity of casings from low density LPGV with various density 1.35-1.60g/sm3 and PGV

reliable data on thermal conductivity of pyrographites, measurements were made on samples processed by standard techniques routinely used for production of standard NFA parts. PGV thermal conductivity does not change monotonously. As temperature rises within the certain range, the thermal conductivity decreases reaching a certain minimum and then rises again (see Fig. 4.24).

The maximum difference between absolute values of heat conductivity for the samples performed at temperatures ~ 1,300 K does not exceed 20 %, while for those made at ~2,400 K it does not exceed 13%, which is quite acceptable for the pyrolytic graphite processed by the same technique. The measurement error X for pyrolytic graphite in all cases does not exceed 8-10 %.

Scatter of thermal conductivity values for standard high-density pyrographite as well as the one of the literature values is attributable primarily to multiple defects, as

well as to micro — and macrostructural differences in PGV parts (see Fig. 4.24) caused by differences in processing history (deposition temperature, heat treatment, gas pressure and flow), sample geometry and other factors. PGN thermal conductivity is typically higher than that of PGV deposited at 2,300-2,500 K (Fig. 4.25). Such change in thermal conductivity with density is due to the structural changes in the material, this assertion being confirmed by the X-ray structural analysis and microstructural characterization.

PGV structure is typically characterized by presence of a significant amount of microcracks (see Fig. 4.26). The reasons for microcracking being (1) stresses gener­ated by temperature gradient across the layer during cooling, and (2) anisotropy of thermal expansion along c and a axes. The thicker the PGV layer, the greater are the stresses and consequently, the higher the propensity to microcracking.

The standard PGV products with wall thickness S of 0.75 mm may feature microc­racks of considerable length that typically pass through multiple growth cones. Upon reaching the cone the crack stops, and the new one is generated in the adjoining lay­ers. Virtually defect-free structure is highly unusual in standard PGV products. As the microcrack plane runs nearly parallel to the deposition surface such microcracks act as large thermal resistances during heat transfer by thermal conductivity in the direction perpendicular to the deposition surface. The number of microcracks has a considerable effect on the PGV thermal conductivity value and on its behavior with temperature change.

Powder metallurgy technique of processing the heat insulating materials has a considerable effect on their structure and properties [10, 27]. Porosity increase in ZrC-based samples from 5 to 70 % leads to thermal conductivity drop by an order of magnitude depending on the structural features, particularly, on skeleton structure factor and closed or open porosity.

For example, porous fibrous ZrC (see Fig.4.27d) processed by thermal diffusion impregnation of carbonized fiber by zirconium show lower thermal conductivity than carbides with continuous skeleton matrix, despite their equal density. Natural structural variation in a porous composition inevitably leads to the scatter in heat conductivity values between samples from the same batch. It is to be noted that the temperature coefficient a thermal conductivity in highly porous carbide materials is similar to the one for the dense carbides, with some influence of heat radiation in pores noticeable only at temperatures above 1,800 °C.

ZrC-based fibers with nearly stoichiometric composition had diameters of 15­20 ^m. Bulk density of carbide, depending on a carbonization regime, varied from 70 to 90 % with thermal conductivity ~2 W/m-K at porosity ~70 %. Compressive strength values of these samples are rather low only 1.2-1.5 times that of tensile strength [33]. At equal porosity levels, porous carbides are stronger than fibrous ZrC, though demonstrating somewhat lower thermal insulating capacity than the latter. Greater shrinkage of fibrous ZrC at temperatures >2,300 K even under moder­ate compressive loads makes porous carbide the material of choice for use in HIPs. Greater structural stability of porous carbide, attributable to its structure, is influ­enced by the choice of processing technique [34]. Zirconium carbide and potassium bromide powders in ratios 1:10 to 1:300 were mixed with binder (natural rubber). This premix was pressed and pre-sintered in vacuum for de-bonding at temperatures above carbide sintering point, with subsequent final sintering to achieve consolidated porous parts. Pre-sintering was effected with heating ramp up to 2,300 K during 11 h, with subsequent isothermal dwell time of 0.5h, whereas final sintering was carried out at heating ramp up to 2,800K within 5h with subsequent dwell time of 0.5h at maximum temperature. This sintering technique yielded products with porosity of 50-85 %, depending on the introduced filler amount. Pore stability at high tempera­tures was achieved through the presence of rather large cellular pores corresponding to the pore-forming agent (filler) particle size.

Formation of special structure of a heterophase composition from alternating carbide-graphite layers (ZrC+Pq) with porosity of 30 % (Fig. 4.27e) deposited from a gas phase get to gain stronger material with thermal insulating properties (Fig. 4.28), which are not conceding to properties of high porous carbides.

It should be note that the temperature coefficient of thermal conductivity of high porosity carbide materials is analogous to that for the dense carbides (Fig. 4.29), some influence of heat radiation in pores are detected only at temperatures above 1,800 °C.

Fig. 4.28 Heat conduc­tivity of compositions ZrC + Pq from alternating carbide-graphite layers:

ZrC 35-40 weight.%, PGV 60-65 weight%, (1) and ZrC 55-60 weight%. PGV 40­45 weight% with a thickness of 0.25-0.5 mm every layer (2)

Fig. 4.29 Heat conductivity of ZrC, П = 6% (1), ZrC +5%C (2), ZrC +5%UC (3) pyrolitic graphite (4), fibrous ZrC carbide p = 70% (5), foam carbide ZrC П = 65 % (6)

Damage of the bearing grid in the course of reactor operation and loss of bearing capacity may lead to an emergency of the reactor. Damage of bearing grids (BG) was revealed at dismantling of channels of reactor IVG-1 after various start-ups in the aspect of linings’ separation from adjoining rows of rods. We remind that the BG are made in the form of sintered block of twisted—four blade rods from ZrC, NbC carbides equimolar composition soldered to lateral surfaces of the block segment facings from pure carbide ZrC.

Though the load-bearing capacity of support grids (SG) was retained, the fact of its fracture confirmed an uncontrollable reliability loss of reactor operation. This neces­sitated the analysis of the stress fields of SG during operation and the need for deeper understanding of possible fracture mechanisms in order to develop recommendations

Table 7.1 Possible strengthening methods for rod-shaped fuel elements

Designated method A о jo (%)

1 Thermal cycling in temperature range from TBP to 2,700K 20

2 Quench hardening by thermal radiation cooling from 2,700K ~30

3 Quench hardening by forced cooling from 2,700K ~60

4 Thermal healing of initial cracks at 2,500K ~40

5 Reactor radiation creation of surface compressive stresses ~80

Combined method 2.4.5 100

for improving SG performance [13]. The results of stress analysis show that SG may be fractured during unstable start stage (before reaching steady-state operation mode) or, under certain conditions, also during cooling stage. When approaching the steady — state regime highest thermal stresses may be reached: up to 400+MPa for tangential tensile stresses on the SG periphery, and up to -70 MPa for compressive stresses near the SG center. Under such stresses, segment linings may crack and delaminate from soldered rod block.

Massive segment linings are proposed to be replaced with thin-walled split-type casing with semi-cylindrical fillers to avoid generation of dangerously high residual stresses. The splitting angle depends primarily on the prerequisite ofretaining the rods in a package during assembly and subsequent processing. Angle value may fluctuate over a wide range: from 5 to ~90° relative to generating line. Implementation of such casing for grid assembly presents no problems, while being also less labor­intensive as compared to lining design. Testing of an experimental batch of proposed SG design has confirmed triple increase of thermal stability of split-type casing SG, as compared to lining design.

Metallographic and X-ray diffraction analysis was widely used in the investiga­tion of materials. The lattice period and phase components in carbide composi­tions were determined by standard methods using URS-50IM and DRON-3 X-ray setups. The chemical and phase compositions of materials were studied by the known instrumental methods of chemical, physicochemical, emission spectroscopy, atomic spectroscopy, and X-ray diffraction analysis. In addition, newly developed methods of porometry,