论文部分内容阅读
鉴于有限元分析耗时耗资源的缺点,为了加速集成电路的互连可靠性分析,提出将传统的有限元建模和人工神经网络(ANN)建模技术结合来实现IC的建模和仿真分析。采用有限元ANSYS参数化设计语言(APDL)实现IC三维模型的自动构建和原子通量散度(AFD)计算,之后通过对计算所得的可靠性数据进行训练和测试,采用神经网络技术对模型的输入输出关系进行建模,使模型达到足够高的精度。神经网络模型构建好之后,可以在短时间产生一个可靠性数据库。通过对数据的统计分析可以得到电路在不同条件下的互连可靠性,进而分析各因素对电路互连可靠性的影响,为集成电路的互连可靠性分析和设计提供重要指导。
In view of the shortcomings of time-consuming and finite-element analysis of resources, in order to speed up the interconnection reliability analysis of integrated circuits, it is proposed to combine traditional finite element modeling and artificial neural network (ANN) modeling techniques to realize IC modeling and simulation analysis . The finite element ANSYS parametric design language (APDL) is used to automatically build the 3D model of the IC and calculate the atomic flux divergence (AFD). Afterwards, the reliability data calculated and calculated are trained and tested. The neural network Input-output relationships are modeled so that the model achieves high enough accuracy. Once the neural network model is built, a reliability database can be generated in a short time. Through the statistical analysis of the data, the interconnection reliability of the circuit under different conditions can be obtained, and the influence of various factors on the interconnection reliability of the circuit can be analyzed, providing important guidance for the reliability analysis and design of the interconnection of the integrated circuit.