The instability characteristics of a dielectric fluid layer heated from below under the influence of a uniform vertical alternating current (AC) electric field is analyzed for different types of electric potential (constant electric potential/ electric current), velocity (rigid/free) and temperature boundary conditions (constant temperature/heat flux or a mixed condition at the upper boundary). The resulting eigenvalue problem is solved numerically using the shooting method for various boundary conditions and the solution is also found in a simple closed form when the perturbation heat flux is zero at the boundaries. The possibility of a more precise control of electrothermal convection (ETC) through various boundary conditions is emphasized. The effect of increasing AC electric Rayleigh number is to hasten while that of Biot number is to delay the onset of ETC. The system is more stable for rigid-rigid boundaries when compared to rigid-free and least stable for free-free boundaries. The change of electric potential boundary condition at the upper boundary from constant electric potential to constant electric current is found to instill more stability on the system. Besides, increase in the AC electric Rayleigh number and the Biot number is to reduce the size of convection cells.
Transverse effective thermal conductivity of the random unidirectional fibre-reinforced composite was studied. The geometry was circular with random patterns formed using random sequential addition method. Composite geometries for different volume fraction and fibre radii were generated and their effective thermal conductivities (ETC) were calculated. Influence of fibre-matrix conductivity ratio on composite ETC was investigated for high and low values. Patterns were described by a set of coordination numbers (CN) and correlations between ETC and CN were constructed. The correlations were compared with available formulae presented in literature. Additionally, symmetry of the conductivity tensor for the studied geometries of fibres was analysed.
The paper deals with fault diagnosis of nonlinear analogue integrated circuits. Soft spot short defects are analysed taking into account variations of the circuit parameters due to physical imperfections as well as self-heating of the chip. A method enabling to detect, locate and estimate the value of a spot defect has been developed. For this purpose an appropriate objective function was minimized using an optimization procedure based on the Fibonacci method. The proposed approach exploits DC measurements in the test phase, performed at a limited number of accessible points. For illustration three numerical examples are given.