低群延迟无源射频线性相位滤波器模块化设计理论研究

In this research project, it is provided that an accurate model of passive RF filter of components, the mutual inductance and a variety of parasitic parameters in the circuit structure; it is determined that the optimization of the complex objective function on low group delay liner phase filter about phase-frequency characteristics and amplitude-frequency characteristics and impedance matching; it is discussed and deduced the first and second order gradient formula about nominal value of objective function of components model, the mutual inductance, and a variety of parasitic. Using computer analysis, it is analysised that the behavior of the objective function, to construct a suitable global optimization method for optimization of the linear phase filter complex objective function. With analysis of the characteristics of the complex objective function, it is used that the fault diagnosis about analog circuit based on modified particle swarm optimization neural network. It is analysised that the factors in the differences between the actual structure and theoretical model of the filter, in addition to the impact factors of working conditions other than the frequency factor, such as deformation, vibration, temperature change. It is deduced to find a design method so that these factors impact on the performance of the filter is fixed. According to the actual structure of the filter model, combined with the frequency range of the filter's applications, and other factors identified former, it is derived the node parameters of the filter structure in batch design by global optimization approach, in order to establish the filter modular design theory.

本项目提出建立无源射频滤波器的元器件、互感系数和电路结构中各种寄生参数的精确模型；确定优化低群延迟线性相位滤波器相频特性、幅频特性和阻抗匹配的复合目标函数；研究推导出目标函数对元器件模型、互感和各种寄生参数标称值的一阶和二阶梯度公式。利用计算机分析和研究上述目标函数的性态，构造一个适合于优化线性相位滤波器复合目标函数的全局优化方法。研究目标函数的特点，运用故障诊断的思想研究基于改进的粒子群算法神经网络的模拟电路故障诊断电路分析。分析滤波器的实际结构模型和理论模型之间所存在差异的因素中，除频率因素影响之外的其他工作条件的影响因素，如形变、振动、温度变化等。研究设计方法使这些因素对滤波器的性能影响为固定。根据滤波器的实际结构模型，结合滤波器应用范围的频率和已确定的其他影响因素，运用全局优化方法推导出滤波器结构批量设计的各节点参数，从而建立滤波器模块化设计理论。

建立了无源射频滤波器的元器件、互感系数和电路结构中各种寄生参数的精确模型；确定优化低群延迟线性相位滤波器相频特性、幅频特性和阻抗匹配的复合目标函数；研究推导出目标函数对元器件模型、互感和各种寄生参数标称值的一阶和二阶梯度公式。利用计算机分析和研究上述目标函数的性态，构造一个适合于优化线性相位滤波器复合目标函数的全局优化方法。研究目标函数的特点，运用故障诊断的思想研究基于改进的粒子群算法神经网络的模拟电路故障诊断电路分析。根据滤波器的实际结构模型，运用全局优化方法推导出滤波器结构批量设计的各节点参数，完成了低、中、高频段三种类型滤波器的设计；编制了调试软件，实现了待调滤波器中元器件、互感和寄生参数的准确计算；试制了线性相位滤波器，用于重要国防建设项目。