RADIATION HEAT TRANSFER IN A SUPERSONIC NOZZLE OF A SOLID-PROPELLANT ROCKET ENGINE
Following from: Radiation heat transfer in a solid-propellant rocket engine
A principal difficulty of radiation heat transfer calculations in a supersonic nozzle is a significant variation of the medium temperature and density along the nozzle axis. In contrast to combustion chamber conditions, the temperature of condensed phase particles in the nozzle is several tens or even hundreds of degrees greater than the temperature of gas, and it is different for particles of various sizes. For this reason, a proper calculation of the radiation heat transfer in a supersonic nozzle is impossible without two-phase flow calculations.
Calculation of concentration and temperature fields of two-phase combustion products in a supersonic nozzle is a complicated physical and mathematical problem due to the substantial velocity and temperature nonequilibrium of the condensed phase. Strictly speaking, one should take into account the influence of particles on the gas flow field, as well as break up and coagulation of particles. Mathematically, the problem is additionally complicated by the 2D flow pattern, particularly in the transonic region, and also by complex particle size distribution of the condensed phase. Both break up and agglomeration of alumina melt particles appear to be important at separate regions of the nozzle flow. The latter makes the problem especially complicated.
The study of gas flow with particles in a supersonic nozzle has a long history. One can remember the early papers by Carlson and Hoglund (1964), Hoffman and Lorenc (1965), Jenkins and Hoglund (1969), and Regan et al. (1971), and the books by Sternin (1974), Pirumov and Roslyakov (1978), Yanenko et al. (1980), and Sternin and Shreiber (1994), which are well known in Russia. Some additional methodological results can be found in other papers (Chang, 1980, 1983; Crowe, 1982; Hwang and Chang, 1988; Carrier et al., 1991; Vasenin et al., 1995). The general methods of two-phase flow calculations have been already developed and one can use the present-day codes for nozzle flow calculations. In this article, taking into account the known difficulties with the use of general algorithms, the flow model considered is as simple as possible. We consider a conical nozzle and suppose that the gas flow field is 1D, and it has spherical symmetry with the center in the cone vertex. Strictly speaking, this assumption is not correct and results in a wrong flow pattern near the critical cross section of the nozzle (Pirumov and Roslyakov, 1978). Nevertheless, the 1D flow analysis seems to be applicable for investigation of the main regularities of heat transfer by radiation in supersonic nozzles of high expansion ratio. It was shown by Dombrovsky (1996a) that the variation of gas flow parameters in a nozzle cross section makes just no difference for the radiation heat transfer. At the same time, it is worth remembering that Klabukov et al. (1988) have used a more detailed flow model in the radiation problem.
The fields of velocity, concentration, and temperature of the condensed phase are determined by calculations of the velocity and temperature delay of single particles. The condition of dynamic and thermal equilibrium between the condensed phase particles and ambient gas is assumed in the initial cross section of the nozzle. The mathematical problem statement is as follows:
where ug(x) and Tg(x) are the known profiles of gas velocity and temperature along the nozzle axis, CD is the drag coefficient, and Nu is the Nusselt number. The values of CD and Nu are determined by using the known semi-empirical relations. Following the paper of Dombrovsky (1996a), consider two high-altitude nozzles with the diameter of critical cross section d* = 0.2 m and 0.02 m, the angle between the nozzle wall generatrix and the axis θ0 = 21 deg, and the expansion ratio ξa = da/d* = 8.7. For brevity, the larger nozzle is hereafter referred to as nozzle 1 and the smaller nozzle as nozzle 2. A comparison of the results of calculations for these nozzles enables us to determine the effect of the nozzle dimension on the heat transfer by radiation. We do not provide the data for the composition and temperature of the combustion products; however, for assessing the level of parameters, note that the temperature in the combustion chamber is about 3600 K, the pressure is about 6 MPa, and the mass fraction of condensed phase particles is about 0.34. These values are typical of solid-propellant rocket engines (Shishkov et al., 1988). In the nozzles considered, the gas temperature near the exit cross section is much lower than the melting temperature of aluminum oxide, Tm = 2320 K. According to Reizer et al. (1973), the crystallization of particles must be included in the case of nozzle 1. It was shown experimentally by Vazhinsky (1980) that the maximum possible relative supercooling of alumina is about 20%, i.e., spontaneous crystallization occurs at the temperature Ts ≈ 0.8Tm. The following expression was given by Henderson (1977) for the velocity of a crystallization wave in a bulk sample:
This formula was recommended for condensed phase particles at Af = 0.72 × 10-6 m/(s·K1.8), nf = 1.8. The value of Lm = 1070 kJ/kg was taken for the latent heat of alumina melting. The results of calculations by taking into account the crystallization kinetics are given in Fig. 1 from books by Dombrovsky (1996b) and Dombrovsky and Baillis (2010). In the graphs, ξ = d/d* is the current expansion ratio, and ρp(ξ) = ρp(ξ)/ρp(1) is the relative mass concentration of the condensed phase particles. A temperature increase during the crystallization of particles is explained by the release of the latent heat of melting. The mass concentration and temperature of particles decrease significantly with ξ. This leads to the corresponding variation in radiative characteristics of combustion products. Note that there is a considerable difference in the temperature of large particles between nozzles 1 and 2.
Figure 1. Variation of particle temperature (a) and relative mass concentration of condensed phase (b) along the nozzle axis: solid lines-nozzle 1; dashed lines--nozzle 2; 1-a= 1 μm, 2-a= 3 μm, 3-a= 5 μm
The spectral absorption coefficient and transport extinction coefficient of two-phase combustion products are calculated as follows:
where αλg is the absorption coefficient of the molecular gas mixture calculated by using the box model (Modest, 2003). Note that this approximation for the absorption coefficient of gases is justified by the predominance of the effects due to the presence of numerous particles. The results of calculations of αλ and βλtr for nozzle 1 are shown in Fig. 2 from the book by Dombrovsky and Baillis (2010). The calculations were performed for the polydisperse condensed phase with gamma distribution of particles by size [see Eq. (10) from the article, “Radiative properties of polydisperse systems of independent particles”] with parameters A = 3 μm-1, B = 2. One can see that αλ drops by about three orders of magnitude when the expansion ratio ξ varies from 1 to ξa. As a result, the relative contribution of the gas phase increases appreciably. The transport extinction coefficient βλtr decreases not as fast as αλ since the value of βλtr is insensitive to the particle temperature. As a result, the transport albedo ωλtr = 1 - αλ/βλtr increases along the nozzle and the scattering considerably dominates over the absorption, especially in the large-scale nozzle.
Figure 2. Variation of the spectral absorption coefficient αλ and transport extinction coefficient βλtr along the axis of nozzle 1 (I) and nozzle 2 (II): 1--λ = 1 μm, 2--λ = 2 μm
In calculations of the thermal radiation transfer in a nozzle, one should take into account the differences between the temperatures of particles of various sizes and the gas temperature. For this reason, the term corresponding to the emission of thermal radiation in the radiative transfer equation is written in the form
where the first term is the condensed phase contribution and the second term corresponds to the gas phase contribution.
The numerical solution for the 2D radiative transfer problem in a nozzle becomes complicated by the following circumstances: a complex shape of the computational region, a sharp variation of parameters along the nozzle in the acceleration region and in the region of c-phase crystallization, a variation of the wall surface temperature, and an indeterminate radiation from the exhaust jet toward the nozzle.
When a lot of condensed phase particles are in the boundary layer on the nozzle wall, the thermal boundary layer may be not quite transparent for thermal radiation. In this case, the problem of the wall radiation flux determination becomes considerably complicated, the more so because the particle temperature in the boundary layer may be much greater than that in the main flow. In this section, we consider only the case of low particle concentration in the boundary layer. Note that this assumption is usually satisfied for the nozzles with a large spread angle, which is used to avoid intense mechanical erosion of the nozzle by impact of condensed phase particles. Under the favorable condition of a transparent thermal boundary layer, the convective and radiative heat transfer do not influence each other, and the total heat flux toward the nozzle wall can be determined as a sum of convective flux and radiation flux from the central region of the flow. It is known that particle concentration in the boundary layer on the nozzle wall may be considerable (Soo, 1989). In this case, one should solve a combined heat transfer problem in the boundary layer.
Generally speaking, thermal radiation affects the temperature of particles in a supersonic nozzle, i.e., the radiation in the axial direction is partially absorbed by comparably cold particles, and they are heated. This effect depends on the particle size. The calculations by Dombrovsky (1996a,b) showed that particles of radius of about 1 μm are heated by radiation in the ξ < 3 zone and cooled by own thermal radiation at ξ > 3. The particles of radius a = 5 μm are cooled by thermal radiation during their motion in the nozzle, even at ξ > 1.7. A variation of the particle temperature under the effect of radiation is small (on a level of 10 K), and this effect can be neglected in practical calculations. It should be noted that the effect of thermal radiation on the temperature of particles may be considerable under off-design conditions of the nozzle flow, i.e., the particles approaching the shock wave are heated by radiation propagating upstream. According to the data by Klabukov et al. (1988), the related variation of the particle temperature may reach 200 K.
To the best of our knowledge, only a few papers are concerned with the radiation heat transfer in rocket nozzles. This may be explained by two reasons. In the first place, the early estimates based on 1D models indicated the insignificant role of the radiation as compared with the convective heat transfer in the nozzle (Pearce, 1978). In the second place, as was mentioned above, a correct calculation of the radiative transfer in a supersonic nozzle is very complicated, i.e., one should calculate the flow field and optical properties of two-phase combustion products and solve the 2D radiative transfer problem in an essentially inhomogeneous medium taking into account a variable temperature of the nozzle surface. Vafin et al. (1981) and Dombrovsky and Barkova (1986) have shown that the previously existing view of the minor importance of thermal radiation for heat transfer in the nozzle of a solid-propellant rocket engine (SPRE) is incorrect, and radiative heat transfer should be taken into account in thermal calculations. Vafin et al. (1981) were probably the first to solve the 2D radiative transfer problem for an SPRE nozzle. The radiative transfer model was based on the P1 approximation. The corresponding boundary-value problem was solved by a finite difference method. Absorption and scattering of radiation by condensed phase particles were taken into account. A series of calculations led to the conclusion about a significant error of a 1D approximation in the supersonic part of the nozzle. At the same time, some of the basic assumptions by Vafin et al. (1981) restrict the possibility of using the obtained results. In particular, the condensed phase was considered as monodisperse, and the thermal and dynamic nonequilibrium of the condensed phase, as well as particle crystallization, were not taken into account. The main, more important, assumption was in ignoring the thermal radiation of the internal surface of the nozzle. Thus, the radiative transfer calculated in the paper by Vafin et al. (1981) may be referred to only at the beginning of the engine running. As shown by Dombrovsky (1996a,b,c), thermal radiation of the nozzle walls is an important factor forming the radiation field in the nozzle of an SPRE. The above-mentioned limitations in the problem statement were removed in papers by Dombrovsky (1996a,c), which are used as a basis of the present section.
An analysis of the radiative heat transfer in a liquid-propellant rocket engine (LPRE) by taking into account the opacity of combustion products due to the presence of soot has been performed by Hammad and Naraghi (1989). The combustion products were considered as a gray medium with isotropic scattering. An additional difference from a similar problem for an SPRE is the cooling of the walls, which makes their radiation insignificant. It was found by Liu and Tiwari (1996) and Nelson (1997) that radiative effects on the wall heat transfer may be significant in chemically reacting nozzles and scramjet combustors. In a recent paper by Byun and Baek (2007), a detailed numerical investigation of radiation heat transfer in an LPRE was presented. The authors used a spectral radiation model with a present-day description of radiative properties of gases and soot particles. It was shown that the effect of radiation (mainly from soot particles) on the heat transfer to the nozzle wall may be significant.
In papers by Dombrovsky and Barkova (1986) and Dombrovsky (1996a), the calculations of 2D radiative transfer in supersonic nozzles were performed by use of the P1 approximation and finite element method. The boundary value problem for the spectral radiation energy density Iλ0 is as follows:
where Dλ = 1/(3βλtr) is the radiation diffusion coefficient, γw is the boundary condition parameter equal to εw/(2 - εw) on the nozzle wall (εw is the wall emissivity), Sw = Bλ(Tw) on the nozzle wall, Tw is the temperature of the nozzle internal surface. The condition of zero radiation flux (γw = 0) was assumed in the critical cross section of the nozzle (at ξ = 1). To determine a correct boundary condition in the exit cross section of the nozzle ξ = ξa, taking into account the small self-emission of a high-altitude jet and radiation backscattering, the computational region was enlarged by conventional initial zone of the jet ξa < ξ < ξ0 with extrapolated optical properties of combustion products. A variation of parameters of the jet initial zone showed that a boundary condition of zero external flux (γw = 1, Sw =0) at ξ = ξa would be crude. This condition leads to additional radiation flux to the nozzle wall due to an error of assumed linear approximation of the angular dependence of radiation intensity in the P1 approximation. As a result, the calculated integral radiation flux from the wall near the nozzle exit may be ~ 30% less than the exact value. As pointed out above, this problem was overcome by insignificant enlarging of the computational region.
The spectral and integral radiation fluxes to the nozzle wall are determined by the following relations:
Following the method described in the article, “Finite element method for radiation diffusion in nonisothermal and nonhomogeneous media,” the boundary-value problem (6) was formulated as a variational problem when the minimum of the functional
should be found. This problem was solved numerically by the finite element method. The solution was obtained in spherical coordinates for the computational region 0<θ<θ0, 1<ξ <ξ0. In analyzing the main results of calculations of heat transfer by radiation in a nozzle, we will restrict ourselves to the computational data for the following two approximations of temperature profiles of the nozzle internal surface:
which correspond to different time moments of the engine running. Some results of calculations for nozzle 1 are shown in Fig. 3 from the books by Dombrovsky (1996) and Dombrovsky and Baillis (2010). One can see that the integral radiation flux is comparable with the convective heat flux in the region of high expansion ratios. The relative role of the radiation may vary severalfold during the nozzle wall heating. The gaseous phase emission and absorption are insignificant, whereas the crystallization of the condensed phase increases appreciably the radiation flux to the wall at ξ > 3. It can be shown that the monodisperse approximation discussed above as applied to the combustion chamber may lead to essential errors in the region of ξ > 3 due to the incorrect determination of the condensed phase temperature.
Figure 3. Convective heat flux qwc and integral radiation flux qw on the wall of nozzle 1: a--k= 1, b--k = 2; 1--calculation ignoring the gas radiation (αλg = 0), 2--calculation ignoring the crystallization of particles (Lm = 0)
To determine the radiation flux to the nozzle wall in a 1D approximation, the assumption of constant gas flow parameters in every cross section of the nozzle was used. When the flow field is considered as central-symmetric, the flow parameters are constant only along the spherical surfaces. With transfer to the axisymmetrical 1D problem, we use the flow parameters near the wall. An error of this approximation is not significant. The radiation flux to the nozzle wall was calculated by using the analytical solution in the P1 approximation. A comparison of the 2D numerical solution with the 1D approximation (see Fig. 4) demonstrates that 2D effects must be included everywhere except for the region adjoining the critical cross section. For nozzle 1, a 1D calculation proves acceptable in the case of ξ < 3, and for nozzle 2, only at ξ < 1.5. The radiative transfer along the nozzle axis leads to a substantial variation of the radiation flux to the wall in the region of high expansion ratios. The examples discussed above are characterized by the predominance of the heat removal due to self-radiation of the wall toward the nozzle exit rather than by heat supply from the high-temperature zone located upstream. Figure 4, reproduced from the book by Dombrovsky and Baillis (2010), further illustrates the important role of the scattering of the radiation by particles. The presence of scattering leads to partial shielding of thermal radiation from the high-temperature region, and to reduction of the radiation flux from the walls toward the nozzle exit cross section. A failure to include radiation results both in incorrect determination of radiation flux and in considerable distortion of the entire pattern of thermal radiation transfer in the nozzle. The hot walls of the nozzle in the region from ξ ≈ 1.5 to 3.5 radiate freely toward the nozzle exit, and the walls in the region ξ > 6-7 are additionally heated by this radiation. The importance of heat transfer by radiation is especially great in the case of small nozzles (see the data for nozzle 2).
Figure 4. Integral radiation flux to the nozzle wall: a--k= 1, b--k = 2; 1--complete two-dimensional calculation, 2--one-dimensional calculation, 3--two-dimensional calculation ignoring scattering
The foregoing inferences about the importance of the 2D effects under conditions of heat transfer by radiation in supersonic nozzles and about the role of scattering are corroborated by the spectral dependences given in Fig. 5 from the book by Dombrovsky and Baillis (2010). In particular, the positions of the extrema of the curves qwλ(λ), calculated in disregard of scattering, is indicative of the radiation source, i.e., in the cross section ξ = 5.8 (Fig. 5a), the small maximum at λ < 1 μm corresponds to radiation from the hot region, while the pronounced maximum at λ = 1.8 μm corresponds to intensive thermal self-radiation of the comparably cold wall of the nozzle. In the exit cross section of the nozzle (Fig. 5b), the predominant effect is that of the radiation received by the wall with a maximum at the wavelength λ = 1.2 μm (that is, from the high-temperature region) and minor uncompensated thermal self-radiation of the wall shows up at λ > 2.6 μm.
Figure 5. Spectral radiation flux to the wall of nozzle 1 for k = 1: 1--complete 2D calculation, 2--1D calculation, 3--2D calculation ignoring scattering
The spectral dependences of the radiation flux shown in Fig. 5 indicate that correct calculations of radiative heat transfer in supersonic nozzles, generally speaking, cannot be performed in a gray approximation. Of course, this does not rule out the possibility of using “gray” engineering models that are oriented to a rough estimate of the integral radiation flux and do not pretend to a high accuracy. The simplest theoretical estimate of the integral radiation flux from the internal surface of a high-altitude nozzle can be obtained by calculating the radiation of the given ring element of the surface to the exit cross section of the nozzle. In so doing, the problem reduces to determining the view factor ψ, and the radiation flux is calculated by the formulas
In the vicinity of the nozzle exit cross section, the expression for the view factor is considerably simplified,
Equations (10) include no dependence on any linear dimensions. Therefore, this approximate solution can correspond to the exact calculations for nozzles of different sizes only on an average. A comparison shown in Fig. 6 indicates that Eq. (10) can be used for estimating the maximum heat removal by radiation from the internal surface of a nozzle with d* > 0.05 m in the range of ξ > 3.5-4, where a 1D approximation would be obviously invalid. Note that the error of such an estimation in calculating the thermal state of the nozzle walls proves rather insignificant because of the nonlinear dependence of the thermal radiation on the surface temperature. The examples of temperature calculations for the nozzle with a titanium wall protected from internal heating by a layer of phenolic carbon and for the carbon-carbon nozzle cooled by radiation from the external surface were considered in the book by Dombrovsky (1996b). It was shown that approximation (10) has an acceptable accuracy, and can be employed for estimates of radiative heat removal from the internal surface of the high-altitude rocket nozzles.
Figure 6. Spectral radiation flux to the wall of nozzle 1 for k = 1: 1--complete 2D calculation, 2--1D calculation, 3--2D calculation ignoring scattering
The results obtained in this study of heat transfer by radiation in supersonic nozzles may seem to be trivial. Therefore, it should be remembered that the calculation of thermal radiation fluxes in the supersonic nozzle by a 1D approximation or disregarding radiation scattering by combustion products, as well as disregarding thermal radiation of the nonisothermal surface of the nozzle, yields physically incorrect results and may lead to considerable errors in determining the temperature of high-expansion-ratio nozzles.
Byun, D. and Baek, S. W., Numerical investigation of combustion with non-gray thermal radiation and soot formation effect in a liquid rocket engine, Int. J. Heat Mass Transfer, vol. 50, no. 3-4, pp. 412-422, 2007.
Carrier, G., Fendell, F., Brent, D., Kimbrough, C., Loucks, S., Hess, E., and Acosta, P., Simple modeling of particle trajectories in solid rocket motors, J. Propulsion Power, vol. 7, no. 2, pp. 185-195, 1991.
Carlson, D. J. and Hoglund, R. F., Particle drag and heat transfer in rocket nozzles, AIAA J., vol. 2, no. 11, pp. 1980-1984, 1964.
Chang, I.-S., One- and two-phase nozzle flows, AIAA J., vol. 18, no. 12, pp. 1455-1461, 1980.
Chang, I.-S., Three-dimensional, two-phase supersonic nozzle flows, AIAA J., vol. 21, no. 5, pp. 671-678, 1983.
Crowe, C. T., Review--Numerical models for dilute gas-particle flows, J. Fluids Eng., vol. 104, no. 3, pp. 297-303, 1982.
Dombrovsky, L. A., A Theoretical investigation of heat transfer by radiation under conditions of two-phase flow in a supersonic nozzle, High Temp., vol. 34, no. 2, pp. 255-262, 1996a.
Dombrovsky, L. A., Radiation Heat Transfer in Disperse Systems, Begell House, New York, and Redding, CT, 1996b.
Dombrovsky, L. A., Approximate methods for calculating radiation heat transfer in dispersed systems, Thermal Eng., vol. 43, no. 3, pp. 235-243, 1996c.
Dombrovsky, L. A. and Baillis, D., Thermal Radiation in Disperse Systems: An Engineering Approach, Begell House, New York and Redding, CT, 2010.
Dombrovsky, L. A. and Barkova, L. G., Solving the two-dimensional problem of thermal-radiation transfer problem in an anisotropically scattering medium using the finite element method, High Temp., vol. 24, no. 4, pp. 585-592, 1986.
Hammad, K. and Naraghi, M., Radiative Heat Transfer in Rocket Thrust Chambers and Nozzles, AIAA Paper No. 1720, 1989.
Henderson, C. B., Effect of crystallization kinetics on rocket performance, AIAA J., vol. 15, no. 4, pp. 600-602, 1977.
Hoffman, J. D. and Lorenc, S. A., A parametric study of gas-particle flows in conical nozzles, AIAA J., vol. 3, no. 1, pp. 103-106, 1965 .
Hwang, C. J. and Chang, G. C., Numerical study of gas-particle flow in a solid rocket nozzle, AIAA J., vol. 26, no. 6, pp. 682-689, 1988.
Jenkins, R. M. and Hoglund, R. F., A unified theory of particle growth in rocket chambers and nozzles, AIAA Paper No. 541, 1969.
Klabukov, V. Ya., Kosolapov, E. A., and Grebenshchikov, L. T., Numerical investigation of high-temperature disperse flows with radiation in convergent-divergent nozzles, Proc. of Int. Heat Mass Transfer Forum, Sect. 1, 2, pp. 153-164, Heat- and Mass Transfer Institute, Minsk, 1988.
Liu, J. and Tiwari, S. N., Radiative heat transfer effects in chemically reacting nozzle flows, J. Thermophys. Heat Transfer, vol. 10, no. 3, pp. 436-444, 1996.
Modest, M. F., Radiative Heat Transfer, 2nd ed., Academic Press, New York, 2003.
Nelson, H. F., Radiative heating in scramjet combustor, J. Thermophys. Heat Transfer, vol. 11, no. 1, pp. 59-64, 1997.
Pearce, B. E., Radiative heat transfer within a solid-propellant rocket motor, J. Spacecraft Rockets, vol. 15, no. 2, pp. 125-128, 1978.
Pirumov, U. G. and Roslyakov, G. S., Gas Flow in Nozzles, Moscow University Press, Moscow, 1978 (in Russian).
Regan, J. F., Thompson, H. D., and Hoglund, R. F., Two-dimensional analysis of transonic gas-particle flows in axisymmetric nozzles, J. Spacecraft Rockets, vol. 8, no. 4, pp. 346-351, 1971.
Reizer, M. Yu., Strunin, V. A., and Manelis, G. B., Possibility of crystallization of condensed combustion products in nozzles, Combust. Explos. Shock Waves, vol. 9, no. 3, pp. 314-319, 1973.
Shishkov, A. A., Panin, S. D., and Rumyantsev, B. V., Working Processes in Solid-Propellant Rocket Engines: A Reference Book, Mashinostroeniye, Moscow, 1988 (in Russian).
Soo, S. L., Particles and Continuum: Multiphase Fluid Dynamics, Hemisphere, New York, 1989.
Sternin, L. E., Fundamentals of Two-Phase Flow Dynamics in Nozzles, Mashinostroeniye, Moscow, 1974 (in Russian).
Sternin, L. E. and Shreiber, A. A., Multiphase Flows of Gas with Particles, Mashinostroeniye, Moscow, 1994 (in Russian).
Vafin, D. B., Dregalin, A. F., and Shigapov, A. B., Radiation of two-phase flows in Laval nozzles, J. Eng. Phys. Thermophys., vol. 41, no. 1, pp. 702-706, 1981.
Vasenin, I. M., Narimanov, R. K., Glazunov, A. A., Kuvshinov, N. E., and Ivanov, V. A., Two-phase flows in the nozzles of solid rocket motors, J. Propul. Power, vol. 11, no. 4, pp. 583-592, 1995.
Vazhinsky, G. A., Experimental determination of crystallization temperature for aluminum oxide particles, Thermal Processes and Properties of Working Media in Engines of Flying Vehicles, Kasan’, Kazan' Aviation Institute, pp. 81-84, 1980 (in Russian).
Yanenko, N. N., Soloukhin, R. I., Papyrin, A. N., and Fomin, V. M., Supersonic Two-Phase Flows with Particle Velocity Nonequilibrium, Nauka, Novosibirsk, 1980 (in Russian).