A Robust Design Optimization Method for Electromagnetic Devices With Interval Uncertainties

High reliability is usually demanded for the design of electromagnetic devices. Interval uncertainties widely exist in the robust optimization problems, and the bounds of the design parameters can be only obtained instead of their specific distributions. To solve this kind of problem efficiently, this paper presents a robust optimization approach based on the Chebyshev interval method to estimate the extreme values of the constraints and objectives. To demonstrate the effectiveness of the method, the Testing Electromagnetic Analysis Methods workshop problem 22 and a brushless dc wheel motor optimization benchmark problem are investigated with interval uncertainties of the relevant parameters. Deterministic and robust optimizations considering the worst case with uniform sampling are also conducted for comparison. The optimization results illustrate the effectiveness of the proposed method.