Remus Banucu

Comparison of a Direct and a Vector Potential Integral Equation Method for the Computation of Eddy Currents

This paper deals with the numerical computation of eddy current problems by means of the volume integral equation method with edge element based shape functions. Two techniques are presented. One allows for the direct computation of the eddy current density. The other one demands the derivation of the electric vector potential to obtain the eddy current density. The differences between these methods are shown, especially the different gauging techniques needed. The efficiency and the accuracy of both methods are examined.

Simulation Based Development of a Valve Actuator for Alternative Drives Using BEM-FEM Code

The importance of natural gas engines is increasing due to low CO2 emissions. To optimize these engines, a direct injection valve has been developed that works with low pressure and is triggered purely electrical. The development of this valve has been supported by numerical simulations using a coupled boundary-element-method-finite-element-method (BEM-FEM) code. This paper deals with the interaction of theoretical considerations, numerical simulations, and measurements on the realized device. In order to achieve the prescribed electromagnetic forces, numerous aspects were considered. The exciting coil, the armature, and the stator geometry is of basic interest. Several concepts have been investigated. Finally, a simple but powerful concept is realized. The simulation results have been compared to measurements on the lately constructed valve.

Design and Optimization of a Contactless Magnetically Levitated Actuator for 4-Axis-Machining

A new design of a machine has been developed which performs contactless actuation in four axis-three translational and one rotational-with high accuracy and controllability in 6-DOF. Remarkable on this concept is the contactless moving part, which enables completely free rotations on the z-axis. The components of the device are optimized especially regarding to the characteristics for micromachining. The optimization algorithm is based on the response surface approximation (RSA) combined with an evolution strategy (ES). The design and optimization process is supported by 3-D numerical simulations using the volume integral equation method (VIEM) coupled with the fast multipole method (FMM) for a fast and efficient computation.

Parallel Hierarchical Block Wavelet Compression for an optimal compression rate of 3-D BEM problems

Todays low cost hardware developments allow for a parallelization of intensive computation processes over several selectable SPU cores by using Intels OpenMP library. But if one applies this feature to a numerical simulation based on a boundary element method (BEM), there is still a huge bottle neck - the shared memory of such a system. So, if a low cost hardware should be effectively used for large BEM problems, a memory compression algorithm that is easily to be scheduled in parallel is of first choice. Within our new idea and development of the Hierarchical Block Wavelet Compression (HWC) which is based on IEEEs JPEG2000 standard for image compression, this bottle neck will be tackled in pure mathematically manners. Furthermore, its parallelization will be discussed and an optimal compression rate for a 3-D electrostatic BEM problem by a rather simple optimization algorithm will be presented.

Automated Optimization in the Design Process of a Magnetically Levitated Table for Machine Tool Applications

A novel magnetically levitated work piece table has been designed. This device is able to perform linear and rotational motions in four axes with high accuracy, so that different operations can be run on the work piece. Numerical simulations support the design of this table. Since the tables motion is idle in a certain range, simulations have to be run for a high number of potential positions. When optimizing the devices layout by varying different parameters, simulations are needed for all positions in which the device shows an extreme behavior. The numerousness of simulations leads to very high efforts concerning pre- and post-processing. Therefore, a new software is developed, which fully automates the whole design process consisting of drawing, meshing, calculating, and optimizing the magnetically levitated table.

Accurate and Efficient Calculation of the Inductance of an Arbitrary-Shaped Coil Using Surface Current Model

In this paper, we present a novel approach for efficient calculation of the inductance of an arbitrary shaped coil with circular cross section using finite element method coupled with an analytical method. To reduce the simulation time and the memory requirement, the surface current model (SCM) instead of the volume current model (VCM) for the conductor modeling is used. According to the magnetic energy formulation and Amperes law, a geometry factor to determine the equivalent height of the SCM for the self-inductance is obtained numerically. To calculate the mutual inductances, further two factors for determining the equivalent distances between adjacent windings are evaluated. Finally, the accuracy and the computational cost of this approach compared with the VCM are presented.

Efficient Compression of 3-D Eddy Current Problems With Integral Formulations

For the calculation of eddy current problems using integral formulations, compression techniques are needed due to the fully populated system matrix. As the system matrix is ill-conditioned, even low compression leads to very high errors and is in most cases unsolvable with classical iterative solvers like CG or GMRES. By using regularization techniques, the condition number is enormously reduced, so that high compression rates can be achieved. In this paper the efficiency of the Block Wavelet Compression combined with the Tikhonov regularization is shown by 3-D eddy current problems. The use of the so called Block Wavelet Compression is presented for the first time for eddy current problems using integral formulations.

Object-Oriented Development of an Optimization Software in Java Using Evolution Strategies

Finding the optimal set of parameters of an often rather difficult system is a major task in numerical optimization. Evolution strategies are optimization techniques based on mutation, reproduction and recombination of configurations as well as environmental selection. A modular composition offers the applicability for different problems. For example, different simulation software packages may be applied and combined with different evaluation strategies. The possibility to approximate the objective function is given. This is facilitated by the use of modern software techniques like design patterns. Therefore, an innovative object-oriented software design for implementing the evolution strategies using Java is presented, discussed and the efficiency of the software package is proven by numerical examples.

Object-oriented development and runtime investigation of 3-D electrostatic FEM problems in pure Java

Nowadays software developments aim on managed code that organizes the memory occupation itself automatically by using the runtime environment. Hence, this is supposed to be very time intensive but simplifies the implementation in an immense way. Therefore, Java is used to determine the advantages and disadvantages of managed code in numerical simulations. In a first step a 3-D electrostatic FEM formulation is implemented in pure Java code by using an object-oriented structure based on design patterns. For an example of a 3-D numerical problem the processing time and the memory occupation are investigated and discussed regarding the use of pure Java code.

Design and optimization of a device with contactless actuation for 4-axis machining

A new design of a machine is developed which performs contactless actuation in four axis - three translational and one rotational - with high accuracy. Remarkable on the concept is the moving part of the machine which is free of contacts, enabling free z-rotation. The components of the device are optimized for different requirements. The optimization algorithm is a combination of an evolution strategy (ES) and an approximation method based on the Response Surface Methodology (RSM). The design and optimization is supported by numerical simulations using the volume integral equation method (VIEM) combined with the fast multipole method (FMM).