Astudy On Optimizations And Designs Of RF Filter

In recent years, with the rapid development of electronic technologies, as an important part of circuits, filters should have higher performances, The conventional design methods for filters may be difficult for high performances. RF(radio frequency) filters can be simulated and optimized by means of electromagnetic simulation software. Filter optimizations using some evolutionary algorithms combined with electromagnetic simulation software tools is recently an attractive research field.The team progress algorithm(TPA) is employed for filter optimizations in this thesis. The external optimizer is used to optimize the return loss of a specific multi-mode ultra-wideband filter. A lower return loss in the passband is obtained.In order to improve several performances of filters, a combinative objective function for the external optimizer is proposed. The combinative objective function with return loss and out-of-band rejection is established to optimize the structures of a ultra-wideband filter. By several tests, the weight coefficients in the objective function is properly determined. The requirement for successful object combination of the objective function is achieved.Furthermore the genetic algorithm(GA) provided by HFSS software is also used for the optimization, the results of which is compared with those of the external optimizer. With the same amount of computation settings, all the optimizations using GA fail, while those using the external optimizer give the better results.The results of the external optimizer is not only better than those of GA, but also consume less time. So it is verified that the external optimizer with the combinative objective function is feasible and it is superior to the GA optimizer built-in the software.