Research On Electromagnetic Optimization Based On Taguchi Method

Taguchi method(TM) is a global optimization algorithm based on orthogonal design.Because of its excellent ability to design the parameters combination, it had been used in many engineering fields. However, the electromagnetic field involves such mathematical and physics problems like complicated model structure, nonlinear objective function and the constraint condition. When Taguchi method was used in solving these electromagnetic problems, level difference would be so small at the last stage of iteration that it could not have the ability to jump out of the local results, and it produced only a single search space in each iteration that the algorithm did not has enough search breadth at the same level difference, which led to an easy losing of global optimal solution. To overcome these drawbacks, there were some improvements proposed in this thesis. The main research works were discussed as follows:(1) This thesis introduced the basic theory of Taguchi method, including the design process of the algorithm, the advantages and disadvantages.(2) Aiming to overcome the shortcomings of Taguchi method, the variation idea of Genetic Algorithm(GA) was introduced here and a Mutation-based Taguchi method(MTM)was proposed. At the same time, according to the annealing theory of Simulated Annealing(SA), an adaptive inner loop mechanism was used not only to diversify search space, but also to control the iteration numbers. There analyzed and compared two ways of variation,and discussed three kinds of mutation operators’ performance when they were used to optimize different test functions.(3) There was design of antenna array pattern synthesis using MTM. The simulation results showed that it can realize low side-lobe as well as nulls in specific directions when optimized the amplitudes of excitation current, its performance could compared to Particle swarm optimization(PSO), but had a less amount of calculation. When optimizing excitation current amplitudes and phase at the same time, the result showed that MTM was effective for such optimization problems.(4) While MTM was applied to design and optimize the multilayer absorber, the electromagnetic parameters and the thickness of the five layers were set as optimizer parameters. Optimized multilayer absorbing material had a lower reflectance and thinner thickness when it compared with TM.(5) Designed PIFA which works in GSM900 by using MTM. Using HFSS to establish PIFA simulation model and using HFSS-Matlab-API as a bridge between Matlab and HFSS to have communication between both kinds of software. System which combined MTM with HFSS could realize the optimization of PIFA’s multi structural parameters. The simulation results showed that the design of PIFA meet the design standard and11 S can reach-54.1dB in the center frequency.