Guangyuan Yang

Electromagnetic Device Design Based on RBF Models and Two New Sequential Optimization Strategies

We present two new strategies for sequential optimization method (SOM) to deal with the optimization design problems of electromagnetic devices. One is a new space reduction strategy; the other is model selection strategy. Meanwhile, radial basis function (RBF) and compactly supported RBF models are investigated to extend the applied model types for SOM. Thereafter, Monte Carlo method is employed to demonstrate the efficiency and superiority of the new space reduction strategy. Five commonly used approximate models are considered for the discussion of model selection strategy. Furthermore, by two TEAM benchmark examples, we can see that SOM with the proposed new strategies and models can significantly speed the optimization design process, and the efficiency of SOM depends a little on the types of approximate models.

A Novel Superposition RBF Collocation Method to Solve Moving Conductor Eddy Current Problems

This paper presents a novel radial basis function (RBF) collocation method to solve the moving conductor eddy current problem. The magnetic field is considered an addition of two fields generated respectively by the excitation current and the eddy current according to the source superposition principle. The corresponding governing equations are decoupled and solved with the RBF. Moving coordinate systems in which the separate fields are computed are also constructed to avoid the model reconfiguration caused by the motion. Electromagnetic field equations are analyzed with kinetic equations and circuit equations together to simulate the motion process. A practical engineering problem is computed to verify the method.

Bayesian Inversion Method and its Information Determination for the Estimation of Particle Size Distribution in Ferrofluids

Using Gaussian numerical integration formula, the problem of estimating the particle size distribution (PSD) in ferrofluids can be converted into an electromagnetic inverse problem. Then we present two Bayesian analytical estimators, minimum mean-square error estimator and maximum a posteriori estimator, to reconstruct the PSD of magnetic particles. In the implementation, weighted minimum norm approach, maximum likelihood estimator, and weighted least square estimator are employed to determine prior information for the unknown parameter. We also present two methods to provide the noise information for the error term. Finally, using Monte Carlo method, we give a ferrofluid example to illustrate the efficiency of the proposed methods.

Multiscale Combined Radial Basis Function Collocation Method for Eddy Currents Analysis in High-Speed Moving Conductors

A novel multiscale combined radial basis function (RBF) collocation method, as a truly meshless method, is presented to overcome the shortage of general RBF collocation method and is applied to analyze eddy currents in high-speed moving conductors in this paper. A typical example is set here to illustrate the accuracy and affectivity of the proposed method, including a comparison with general RBF collocation method and finite element method (FEM).

Using Improved Domain Decomposition Method and Radial Basis Functions to Determine Electromagnetic Fields Near Material Interfaces

This paper presents a novel approach for solving problems of electromagnetic fields near material interfaces. This novel approach is based on combination of radial basis functions (RBF) and improved domain decomposition (DD) method. We discuss three radial basis functions in this condition, compared with common domain decomposition method in two subregions example. To get the same precision, improved method has extended the range of parameter C in radial basis functions. We also apply the method to the analysis of heterogeneous multi-subregions example, a turbo generator rotor winding. The simulation results demonstrate good agreement with practice, showing the adequacy of the proposed methodology for electromagnetic analysis.

Domain Decomposition Combined Radial Basis Function Collocation Method to Solve Transient Eddy Current Magnetic Problems With Moving Conductors

Radial basis function (RBF) collocation method is a kind of pure meshless numerical method and has been applied to solve static and transient electromagnetic problems. Especially, it shows a great advantage in the computation of moving conductor eddy current magnetic problems. To simulate the conductor movement, the field equations are decoupled with superposition principle and solved by time-domain iteration under moving coordinate systems. One problem is that the coefficient matrix of RBF governing equations, which needs to be computed in each iteration step, is full. As the number of RBF nodes increases, the computational capacity will grow rapidly. The domain decomposition method (DDM), which divides the solving domain into several subdomains, could be conveniently combined with RBF collocation method. This paper first applies DDM combined RBF collocation method to compute transient eddy current magnetic field problems with moving conductors. With this novel method, the iteration only proceeds in the subdomains containing conductors. And the magnetic field in the subdomains without conductors needs to be computed just once before the iteration. The dimension of the coefficient matrix computed in the iteration is only determined by the number of nodes in the corresponding subdomains and on the interfaces. An engineering problem is computed to show that the DDM combined RBF collocation method is much more efficient than the normal one.

Linear unbiased estimators for particle size distribution of magnetic nanoparticles

We present two statistical analytical approaches to estimate the particle size distribution of magnetic nanoparticles. They are termed best linear unbiased estimator and linear minimum mean square error estimator. These approaches are implemented and quantified within the formalism of linear unbiased estimation theory. To illustrate the efficiency of the proposed approaches, we give two examples of the application to the particle size distribution analysis in ferrofluids with normal and lognormal samples. In both cases we compare the reconstruction distributions using our methods with those calculated via the electron microscopy images of the ferrofluid particles.

An Improved Population-Based Incremental Learning Method for Objects Buried in Planar Layered Media

An evolutionary algorithm, the estimation of distribution algorithm (EDA), is used to reconstruct the objects that buried in planar layered media. It is essential that fast forward solvers be used to solve the forward scattering problem for the nonlinear inverse scattering methods, since it can avoid errors by approximation. The EDA is a predominant all-round optimizing method in the macroscopic simulation of evolution process species of nature. Recent studies have shown that the EDA provides better solution for nonlinear problems than the microscopic evolutionary algorithm, such as genetic algorithm (GA) in some cases. The EDA is simpler, both computationally and theoretically, than the GA. We discuss how this can be used to calculate the permittivity and conductivity of the targets. We show preliminary results indicating the potential of reconstruction for buried objects. Compared with other methods, the experiment result shows that the EDA algorithm reduces the number of iteration.

Sequential design of experiments techniques for the optimization design of electromagnetic devices

Three sequential design of experiments (uniform, Latin hypercube and central composite designs) techniques are presented to deal with the optimization design problems of electromagnetic devices. These methods can be employed not only as the sampling techniques for the sequential optimization method, but also as a direct sampling optimization method. Thereafter, we use a TEAM Workshop problem to demonstrate the efficiency of the proposed methods. From the experiments and the comparison, we can see that the proposed methods can produce satisfactory solutions, and the compute cost of the total finite element analysis can be remarkably reduced.

Application of meshless collocation method to solve eddy current magnetic field problems involving conductor movement

This paper applies the Radial Basis function (RBF) collocation method to compute the eddy current magnetic field coupled with conductor movement. Because of the linear characteristic of the RBF collocation matrix equation, the unknown magnetic vector potential could be easily decoupled and directly expressed in moving coordinate systems without model recast. This offers a great convenience to deal with the conductor movement. Time-domain iteration is constructed to solve the governing equations. And in each iteration step, the magnetic field and kinetic equations are calculated alternately. Two numerical examples, the TEAM workshop problem 28 and a transient coil gun problem are considered to analyze the accuracy and efficiency of this method.