The residual neural network is used to carry out machine learning on the steady-state flow field of the hemisphere-on-cylinder laser turret model in the range of Ma=0.3~0.8, and the subsonic/transonic flow field under any incoming flow conditions in this range is established. The prediction accuracy of this model is evaluated for beam wavefront distortion under different view-of-field angles. The learning model reproduces flow characteristics such as boundary layers, flow separation, and separated shear layers in turret flows, including in particular unanchored shock discontinuities in transonic flow. The wavefront distribution based on the predicted flow field under different viewing angles is basically consistent with that calculated based on the flow field of CFD. This machine learning method provides a strategy for adaptive correction of laser turret aero-optical effects in the engineering field.

Based on HF(Hartree-Fock) calculation method with 3-21G basis set, we discussed structural and dissociation characteristics of C₃H₄O gas molecules under different external fields (from -10.28 V·nm^-1 to 10.28 V·nm^-1). It is found that as the electric field increases along the direction of molecular conjugated single bond the total energy increases. The bond length of C-C double bond and C-C single bond decreases. The bond length of C-O double bond increases. The dipole moment decreases. Energy gap E_G increases. The infrared absorption peak has both red shift and blue shift at different frequencies and the intensity of IR also changes. The molecular dissociation performance is as follows: The potential energy barrier decreases with the increase of the external electric field. As the electric field reaches 25.71 V·nm^-1, the potential energy barrier almost disappears and the dissociation energy decreases gradually with the increase of the electric field, which indicating that the dissociation difficulty of C₃H₄O gas molecules under electric field decreases gradually. It provides reference for the study of dissociation characteristics of C₃H₄O gas molecules or mixtures containing C₃H₄O in external electric field.

According to the fractal scaling law of microstructures, the particle and pore structures of random porous media were reconstructed with Monte Carlo method. The seepage property of multiscale porous media was studied with a fractal capillary bundle model. A quantitative relation between macroscopic seepage properties and microscopic structures was addressed. It shows that microstructures of porous media reconstructed by the fractal Monte Carlo method are close to that of real media. Calculated results for gas flow through porous media agree well with those by lattice Boltzmann method. It is shown that gas permeability increases with the increment of Knudsen number. The pore fractal dimension takes important effect on the microscale effect of gas flow through porous media, while the influence of tortuosity fractal dimension on the ratio of apparent gas permeability to intrinsic permeability is marginal. The method shows advantages that the convergence speed is fast and the calculation error is independent of dimension. It is helpful in understanding seepage mechanisms of multiscale porous media.

Based on self-developped parallel software platform PANDA, algorithm design and parallel implementation of multipoint-base-excited harmonic response were carried out via modal superposition method. Corresponding parallel solution module was constructed, rightness of which is validated by comparisons with commercial FE software. Parallel module was successfully applied to harmonic response analysis of SG facility. An effective parallel solution of extensive computing scale up to 1.188 billion DOFs was achieved. It shows that parallel solution module is suitable to large-scale refined harmonic response analysis of complex equipments, behaving a strong parallel computing scalability up to thousands of CPU processors, and going far beyond the scale of commercial FE software.

From mass conservation and energy conservation of hotspot, dynamics of hotspot combustion involving shell-mix are constructed by recalculating special internal energy and heat capacity, in which bremsstrahlung radiation loss and other energy transfer are considered. With power balance in hotspot, relations between fractions of shell-mixing in hotspot and ignition threshold and self-sustainable combustion are investigated. Theoretical analysis and numerical simulation show that enhancement of radiation loss induced by shell-mixing plays dominated role in failure of ignition. By adjusting materials and mixing fraction, effects to combustion paths are studied. Finally, two methods for degrading impact of shell-mixing are proposed.