New soft magnetic materials made possible the use of the magnetic amplifier technology in designing competitive electric power supplies. This technology is used in the Swedish. fighter aircraft Gripen, being also attractive for future more electrical aircraft systems due to the possibility to achieve a compact and robust design. A modeling approach of a magnetic amplifier based on the magnetic hysteresis of the core material is presented here for a common amorphous magnetic alloy.
The development of HVDC converters, local generation and tractions systems increases the demand for power converters. While a lot of studies have been conducted on the power electronics, there are still open questions concerning the contained power transformer that operates at a few kilohertz. As power and loss increase with frequency, a smaller transformer can be constructed supposing cooling is increased or losses are reduced.
New modeling approaches have been studied and compared with existing models. This concerns the electrical, magnetic and thermal behavior of the transformer. Common transformer models are based on inductances with effective permeabilities. Such mod-els cannot use hysteresis models. A new transformer model based on the lumped ele-ment approach has been derived. The model combines magnetic hysteresis, eddy currents and thermal diffusion such that their interaction is simulated correctly. The model has been optimized with respect to accuracy and the required number of lumped elements. A field homogenization procedure for a magnetic lamination with eddy currents is presented. The previously suggested magnetic hysteresis model has been further developed and numerically optimized and the agreement with the one-dimensional measurements increased. The overall transformer model was run for a realistic geometry. A novel 2D measurement set-up for characterization of magnetic sheets has been designed and constructed to operate up to a few kilohertz. The ex-perimental verification of the set-up proves that the magnetic field amplitude and es-pecially homogeneity increased considerably. The latter allows larger pick-up coils. The set-up is much smaller than the previous and allows small samples, which is important in the medium frequency range, where a core material often is manufactured on small rolls. In addition, the necessary electric excitation decreased substantially. Measure-ments are required to characterize and evaluate prospective transformer core materials such that they can be used as a material database for a simulator program. A measurement program including automatic demagnetization has been written. One- and two-dimensional magnetic measurements on nonoriented and grain-oriented materials were performed. Rotational power loss was measured for different axis ratio of the rotating magnetic field. The resulting hysteresis loops were explained. The hysteresis model has been extended for isotropic materials and the result agrees with the two-dimensional measurements.
Rotational power loss was measured on non-oriented and grain-oriented silicon steel sheet under different rotational field conditions. A novel two-dimensional measurement system was used. It consists of two perpendicular symmetrical yoke pairs arranged such that the magnetic field is highly homogeneous. The cross product of the H- and the B-vector was adopted as the correct estimation of the rotational power loss rather than the sum of the alternating losses in x-and y-direction. This approach was justified by means of measurements. A number of results are similar to the observations made earlier with rotational single sheet tester (RSST). However, no hard direction was observed for the non-oriented material, and the common RSST is questioned. Controlling H instead of B worked well.
An equation system for the anhysteretic curve and the distribution function required by a hysteresis model is suggested. A least-square fitting confirms that the solution is unique and gives optimal agreement with the measurements. Moreover, the used fitting function with three parameters replaces the measured vector, which increases simulation performance.
physical Cauer circuit is used to model skin effect in a conductor or eddy currents in a magnetic lamination. A physical lumped element model was chosen, because equivalent circuits are inconvenient for nonlinear materials such as magnetic materials. Approaches to discretize the used circuit are presented and compared. An iterative fitting technique is suggested that requires fewer elements and yields higher accuracy than an exponential discretization.
A transformer model is presented where both winding and core were modeled by Cauer circuits in order to take into account eddy currents. The physical property of the magnetic circuit allowed the insertion of a hysteresis model. Thermal effects in the winding insulation were also modeled by a Cauer circuit. The discretization of the circuits was optimized by an iterative method that fits the reluctance of a core part or the impedance of a winding layer to known analytical solutions. This reduced the number of required circuit elements substantially and the accuracy was increased.
A novel two-dimensional measurement set-up for magnetic sheet characterization has been designed and constructed. Experimental verification shows that the inhomogeneity of the magnetic flux through the sample was smaller than the measurement error, whereas the required electric excitation decreased considerably. The aim of this paper is to prove that the new measurement system is superior to the Epstein frame and the rotational single sheet tester (RSST) and can be recommended.
Recently it has been shown that a physical hysteresis model has a unique mathematical solution for a given magnetic material. This solution gives maximum agreement between measured and simulated virgin curve and major loop. As the analytical solution is complicated and difficult to find it was necessary to find the optimum by an approximation. Standard optimization methods found only local optima, because the system is non-linear and non-convex. This paper shows how a stochastic method can be applied to fast approach the global solution.
Modelling the whole power system of an airborne system requires very high computational capacity. This paper presents a possible approach to overcome this obstacle when considering new technology and topologies in airborne systems. Using controlled current converters and electromechanical actuators for rudder control are examples of such new technology. The governing idea is to separate the high (electrical) and low (mechanical) frequency parts [1] of the model and thereafter only consider the low frequency part. Electrical components as current and/or voltage converters/inverters are then desccribed in terms of efficiencies or in circuit diagrams in form of load dependent resistors [2] and in electromachines by means of load dependent loss resistances equivalent with their electromechanical efficiencies [3]. Step-down and step-up voltage converters are desccribed as fictive DC/DC transformers. This approach makes it possible to reduce the complexity of power system models of tentative systems to such extent that the resulting computational tool can be used for studies of the system performance during entire flight missions and/or for optimisation.