Search

The miniaturization of electron devices has its advantages in dense circuits and high frequency devices. As can be seen in the current literature, considerable effort is devoted on the problem of miniaturizing modern electronic devices to allow high speed and high density. Considering the effects in short channel metal oxide semiconductor field transistors (MOSFET), in practice the device is scaled to preserve the long-channel characteristics after miniaturization. In this work, several parameters are chosen and their sensitivity analysis is used for the miniaturization of a MOSFET. The objective is to mininturize the device in such a way to avoid the break down effects at high bias.

The shape optimization of coupled systems by a gradients-based method is presented. The design specifications are in one system, while the critical parameters are in both systems. The method is demonstrated using an induction heating system. The magnetic and thermal models coexist in the same geometry. The eddy currents calculated from the electromagnetic solution is used as the thermal sources for the thermal finite element analysis. The objective is to achieve a required steady state temperature profile by modifying the geometry of the magneto-thermal domain. The objective function, defined as a function of the state variable temperature is no longer linked directly to the design parameters of the magneto-thermal system through the classic design sensitivity analysis but through the 'coupled' one. The proposed algorithm allows the calculation of the gradient of the object function with respect to the design parameters.

In this paper we present the use of electromagnetic signal processing for tracking and estimation. The computation of the backscattered electromagnetic field is given for the specific case of a simplified aircraft model. An on-line electromagnetic estimator that estimates the orientation of the aircraft from the field information is derived. We also give the overall signal processor which includes the electromagnetic sensor and the tracker. Similar signal processors can be developed for many other systems.

Nondestructive testing (NDT) is used to identify the anomalies and defects in inaccessible locations. Various techniques of optimization are used in NDT. In this work, the Artificial Neural Networks (ANNs) are applied with NDT to identify a crack in a conducting medium. In general, deterministic techniques are used with the back propagation algorithm (BP) to train the neural networks. The ANNs which are trained by a deterministic method have a tendency to get trapped in local minima. In this paper a stochastic version of the gradient descent is applied to train the ANNs and it overcomes the difficulties of local minima caused by the sinusoidal fields. The stochastic version used in this approach is based on the Metropolis algorithm which is frequently used in the simulated annealing.

The optimization of electromagnetic devices governed by the Poisson equation is now well established. This paper extends the methodology to scattering problems. In particular, the permittivity and permeability of the coating on an aircraft fuselage and its thickness are computed so as to minimize the scattered signature. It is shown that the reflection coefficient has multiple minima and maxima as the thickness of the coating is varied, thereby making search techniques superior to the usually efficient gradient methods. On the other hand, as the permeability and permittivity are varied, the reflection coefficient varies gradually to its one minimum, except for a few local minima, thereby making gradient techniques very powerful.

Symbolic knowledge representation has made possible the representation of a broader range of engineering knowledge in order to provide new kinds of computer support for both analysis and design. This paper outlines electromagnetic field computation knowledge and reviews the roles of numerical, geometric, and symbolic representations of such knowledge. Some comments on integrating electromagnetic device design knowledge are also offered.

This paper presents the shape optimization of windings to minimize the eddy current losses. The method used for this optimization combines the finite element analysis of the time harmonic electromagnetic field and gradient-based numerical optimization techniques. The objective of this type of optimization problem is to limit the losses by modifying the shape of the windings.

The aim of this paper is twofold: (a) to extend the design sensitivity analysis of finite element eddy current formulations to the situation of constraining the total induced current by applying Ampere's law and (b) to present a formulation for the minimum weight optimization of magnetic shields. In the first step an existing finite element formulation is implemented and validated with test measurements. In the second step design sensitivity analysis is derived for this formulation. The formulation of the optimal design of a shield is based on these sensitivity calculations and solved by nonlinear programming methodologies. The presented formulation results in a large number of explicit and nonlinear constraints, which require the utilization of appropriate nonlinear programming methods. The presented optimization examples demonstrate that grounded shields are far more effective than ungrounded ones.