Robust design of high field magnets through Monte Carlo analysis

Abstract

An effective approach to the optimal design of electromagnetic devices should take into account the effect of mechanical tolerances on the actual devices performance. A possible approach could be to match a Pareto optimality study with a Monte Carlo analysis by randomly varying the constructive parameters. In this paper it is shown how such an analysis can be used to allow an expert designer to select among different Pareto optimal designs.

An Efficient Tabu Search Algorithm for Robust Solutions of Electromagnetic Design Problems

“…Therefore, the preferable design solution is not the globally optimal solution, but the one that is close to the optimal solution while having a high tolerance or robustness to small variations in the decision parameters. In this regard, there are increasing endeavors devoted to the study of robust design techniques [1], [2]. Nevertheless, studies on robust design techniques have a short history of less than ten years, and are still in their infancy.…”

“…Since there is no closed-form objective function for an inverse problem, the expected fitness function of a solution is generally determined from (2) where is the number of sample points generated in the neighborhood of the specific point . From (2) it is obvious that to determine the expected fitness value of a specific solution, a set of additional function evaluations are required.…”

“…Since there is no closed-form objective function for an inverse problem, the expected fitness function of a solution is generally determined from (2) where is the number of sample points generated in the neighborhood of the specific point . From (2) it is obvious that to determine the expected fitness value of a specific solution, a set of additional function evaluations are required. Considering the fact that thousands of iterations are generally required for a global optimizer to converge to the final solution and fitness evaluations are expensive because of the common usage of finite-element analysis, the determination of the expected fitness using (2) is clearly not viable for many engineering design problems.…”

A New Sensitivity-Based Reliability Calculation Algorithm in the Optimal Design of Electromagnetic Devices

“…The MCS method, a direct samplingbased approach, becomes computationally expensive especially when the constraint functions are to be calculated by using numerical methods such as the finite element method (FEM). It is because the MCS method requires as many sampling points as possible (like a few millions) to guarantee its accuracy [6]. To alleviate its expensive cost, the MCS method is normally applied to a meta-model [7] constructed by the Kriging or response surface method, however, until now, for reliability calculation, there is scarcely any researches about combination of the MCS and sensitivity analysis assisted by the FEM.…”

Reliability Assessment on Different Designs of a SMES System Based on the Reliability Index Approach

“…Until now, only Monte Carlo simulation (MCS) has been used in electromagnetic design problems to evaluate quantitatively the statistical property of the performance function of interest, considering the random variables [1][2][3][4]. However, the method has a major drawback: huge computation time is required for its numerical implementation because there are many function calls, although the probability information of random variables is given.…”