Bi-layer Mo films were deposited on soda-lime glass substrates using DC magnetron supttering. Increasing gas pressure, the resistivity varied from $1\times10^{-5}\; to\; 8.3\times10^{-3}\; \Omega.cm$. Furthermore, stress direction yielded compressive-to-tensile transition stress curves. The micro-structure of the compressively-stressed film which had poor adhesion consists of tightly packed columns, but of the tensile-stressed films had less dense structure. Under all gas pressure conditions, Mo films exhibited distinctly increasing optical reflection with decreasing gas pressure. The expansion of (110) peak width with the gas pressure meant the worse crystalline growth. The impurity levels in the Mo film exhibited highly concentrated Na, Se and O elements due to less dense micro-structure. The degree of Na diffusion depends on the type of the glass substrate used and the nature of the Mo film.

The piezoelectric properties and the doping effect for$0.95Pb(Zr_xTi_{1-x})O_3+0.05Pb(Mn1/3Nb2/3)O_3$ compositions were studied. Also, the heat generation and the change of electromechanical characteristics, the important problem in practical usage, were investigated under high electric field driving. As a experiment results under low electric field, the values of kp and $\varepsilon33T$ were maximized, but Qm was minimized(Kp=0.57, Qm=1550) in the composition of x=0.51. In order to increase the values of Qm $Nb_2O_5$ was used as a dopant. As the result of that, the grain size was suppressed and the uniformity of grain was improved. Also, the values of kp decreased, and the values of Qm increased with doping concentration of $Nb_2O_5$. As a experiment results under high electric field driving, when vibration velocity was lower than 0.6[m/s], the temperature increase was $20[^{\circ}C]$, and the change ratio of mechanical quality factor was less than 10[%]. So, its electromechanical characteristics was very stable. Conclusively, piezoelectric ceramic composition investigated at this paper is suitable for application to high power piezoelectric devices.

The piezoelectric properties and the doping effect for $0.95Pb(Zr_xTi_{l-x})O_3+0.O5Pb(Mn_{1/3}Nb_{2/3})O_3$compositions were studied. Also, the heat generation and the change of electromechanical characteristics, the important problem in practical usage, were investigated under high electric field driving. As a experiment results under low electric field, the value of $k_p$ and ${\varepsilon}_{33}^T$ were maximized, but $Q_m$ was minimized $(k_p=0.57, Q_m=1550)$ in the composition of x=0.51. In order to increase the values of $Q_m$, $Nb_2O_5$ was used as a dopant. As the result of that, the grain size was suppressed and the uniformity of grain was improved. Also, the values of $k_p$ decreased, and the values of $Q_m$ increased with doping concentration of $Nb_2O_5$ . As a experiment results under high electric field driving, when vibration velocity was ower than 0.6[m/s], the temperature increase was 20[℃], and the change ratio of mechanical quality factor was less than 10[%]. So, its electromechanical characteristics was very stable. Conclusively, piezoelectric ceramic composition investigated at this paper is suitable for application to high power piezoelectric devices.

A simple methode for calculating the forward voltage drop of IGBTs is presented, on the voltage drops on the p+ body, the reverse biased depletion region between p+body and epi-layer, the epi layer, and the forward biased collector junction. The decrease of the total current density in the epi layer near the p+ body is taken into account. The proposed methode allows a simple but accurate determination of the forward voltage drop in IGBTs, avoiding the complex path taken in the previous model for the forward voltage drops on channel, accumulation region, and epi region. Numerical simulations for 1kV NPT-IGBT with a uniformly doped collector are shown to support the analytical results.

Improvement on the switching characteristic of IGBT by means of the uniform and local lifetime control is studied numerically using two-dimensional simulator, MEDICI. In the case of uniform lifetime control, the on-state and switching characteristics are simulated as a function of lifetime, and compared with the experimental results reported, which allows a relationship between dose of electron irradiation and controlled lifetime. In the case of local lifetime control, simulations are carried out by varying the position, width, and lifetime of the locally controlled region, and the results are compared with the characteristics for the case of the uniform lifetime control. The turn-off time of the device with an optimized locally controlled region is found to decrease from about $4.5\mus$ to 0.11$mutextrm{s}$ while the forward voltage drop increases from 1.37V to 2.61V in comparison with that for the uniform lifetime control.

The partial discharge(PD) signal, may seems to be stochastic and merely random, was investigated using the method to discern between chaos and random signal, e.g. correlation integral, Lyapunov characteristic exponents and etc. For the purpose of obtaining experimental data, partial discharge detecting system via computer aided acoustic sensor, detect PD signal from the insulating system, was used. While this method is very different from typical statistical analysis from the point of view of a nonlinear analysis, it can provide better interpretable criterion according to the time evolution with a degradation process in the same type insulating system.

In this study, the characteristics of various electricity and high resistivity ash collection were measured at superimposed nanosecond pulse energization. The pulse was generated through the hydrogen thyratron. Superimposed pulse was nanosecond duration pulse that had fast pulse rising time 100[ns], short pulse width 500∼850[ns] and pulse fall time 400∼750[ns]. The results of this study are summarized as followings; Frist, compared with the DC ESP(electrostatic precipitator), superimposed pulse ESP could induce higher peak voltage, and had the about 30[%] reduction of ESP collection area and 50∼80[%] reduction of power consumption. showing the same efficiency. Second, enhancement factor H had 1.9∼3.6 with the increase of migration velocity.

In this paper, partial discharge (PD) measurement was performed to evaluate the quality of the cable joint and termination constructions. The resistive coupling technique for PD detection using resistivity of semiconducting layer of the cable in the accessories, such as joints and terminations. With high frequency PD (HEPD) measurement, an excellent sensitivity below 20pC could be achieved under unshielded condition. The localization of the defects in the accessories could be identified. During heating cycle, PDs were monitored and analyzed. At that time, the PDs were dependent on the temperature of the heating cycle and showed cyclic behaviors, which were attributed to local delamination of the interfaces, between epoxy unit and stress relief cone(SRC) and between SRC and cable, due to the difference of thermal expansion. As a conclusion, HFPD measurement technique was proven to be an effective diagnostic method for qualification of extra high voltage (EHV) cable accessories. With this technique, the optimal design of the components of the accessories could be verified not only in an early stage but also under operating condition. This technique would result in the improvement of the reliability of the EHV cable accessories.

It is very important not only to detect PD signal but also to locate PD source in power cable line including its accessories. In this paper, the PID localization technique using interruption of metal sheath of cable or accessory is described. By interruption of metal sheath, the polarity of PD pulse to be measured on each part can be compared. With this technique, the exact location of PD in the accessory which has some defects can be found. Although the 'Soft' interruption method, that is, cable sheath is divided by semi-conductive layer with the resistance of higher than 1 k$\Omega$, was used instead of 'Hard' interruption method, it is also possible to detect exact location of PD. This technique is considered to be very effective for detecting PD localization in the type test of cable and accessories.