微波滤波器的装调融合建模与反演

微波滤波器作为关键的选频装置，在设计制造周期中存在一些突出问题，如设计制造周期数据割裂以及数据缺乏有效利用，从而导致设计制造阶段脱节，降低了设计制造迭代效率。文中提出了一种微波滤波器装调融合建模与反演方法，根据滤波器的设计信息、制造数据以及滤波器机理采用自顶向下的方式构建滤波器知识图谱，利用知识图谱对微波滤波器的数据与机理进行规范化存储；利用机理数据融合建模的方法构建微波滤波器正向演进模型；基于正向演进模型，结合渐进空间映射算法，构建微波滤波器快速装调反演模型。演进与反演实验结果表明，预测电性能与实际电性能基本重合，故障滤波器通过两次装调即可满足设计指标。该方法能够实现微波滤波器设计制造数据的有效关联、微波滤波器性能演进的有效预测以及快速装调反演，提高了设计制造周期数据的利用率，增强了滤波器设计与制造的协同。

Fusion Modeling and Inversion of Microwave Filter Assembly

As a key frequency selection device, microwave filters have some prominent problems in the design and manufacture cycle, such as the data fragmentation of the design and manufacture cycle and the lack of effective use of data, which leads to the disconnection of the design and manufacture stages and reduces the iterative efficiency of the design and manufacture. In this paper, a fusion modeling and inversion method for the assembly of microwave filters is proposed. According to the design information, manufacturing data and mechanism of the filter, the filter knowledge map is constructed using top-down method, and the data and mechanism of microwave filters are stored in a standardized way using the knowledge map. The forward evolution model of microwave filters is constructed using the method of mechanism data fusion modeling. Based on the forward evolution model, the rapid assembly inversion model of microwave filters is constructed by combining the progressive space mapping algorithm. The evolution and inversion experiment results show that the predicted electrical performance is basically overlapped with the actual electrical performance, and the faulty filter can meet the design requirements just after two assemblies. This method can achieve the effective association of the design and manufacturing data, effective prediction of microwave filter performance evolution and rapid assembly inversion, which improves the utilization rate of the data of design and manufacture cycle, and enhances the collaboration of filter design and manufacturing.