Dissertation > Excellent graduate degree dissertation topics show

Study on Optimal Design Method for Micro-electronic Packaging Device Based on Artificial Neural Network

Author: CaiMiao
Tutor: YangDaoGuo
School: Guilin University of Electronic Science and Technology
Course: Mechanical and Electronic Engineering
Keywords: Microelectronic packaging technology Optimal Design Interface spallation Back-propagation neural network ( BPNN ) Principal component analysis Genetic Algorithms Integrated analysis Optimal combination
CLC: TN402
Type: Master's thesis
Year: 2009
Downloads: 28
Quote: 2
Read: Download Dissertation

Abstract

The crack of the interface layer is one of the main failure mode of the microelectronics packaging device. International Technology Roadmap for Semiconductors (ITRS) In recent years, the report stressed that a better understanding of the behavior of the interface and the key to be able to characterize and control the interfacial strength as a future chip development and manufacturing. Interface spallation affected by the damp heat stress, more and steam pressure, integrated wet, heat and steam pressure of interface reliability problem is particularly critical. Secondly, the study of the reliability problems purpose is to improve the reliability of the device, which is the reliability of the design of one of the key. Therefore, it is necessary to explore the optimal design of the device parameters on the basis of interfacial spallation failure. In this paper, the microelectronics packaging device interface layer crack failures, in-depth study of the optimal design method based on improved BP neural network package devices. The main contents are as follows: 1) on the basis of existing technologies for microelectronics packaging device interface layer crack failures, improve and proposed methods of analysis of an integrated heat, humidity and vapor pressure. On the one hand its feasibility; the other hand, preliminary exploration plastic device parameters Optimization, and demonstrate the need to explore a packaged devices suitable for optimization design method. 2) the use of molded sealing material (EMC) fatigue test data, in-depth study of the improved method of BP neural network based on principal component analysis and genetic algorithm, and proved its stability and practicality. 3) improved neural network based on in-depth study of the packaged device optimization design method. Improved neural network model, a packaged devices based on improved neural network optimization design method, and to traditional plastic DR-QFN devices in the optimal design of the thermal - mechanical action, for example, discussed in detail the optimal design process, analyze problems with packaged devices design materials with dimensions. 4) programming optimization design method and the initial realization of the user software interface, described in detail the overall design process to optimize the design of software systems, and gives the key block code function, parameter optimization process automation and user-oriented the design of the software interface. 5) for the interface spall failures, integrated moisture, heat, and steam pressure, to achieve the optimal design of a new type of plastic QFN devices, and made a presentation of the input parameters of the program is running. 6) Failure Cases for the production site, based on the analysis of the binding assay to further verify the feasibility of the optimization design method. The results show that: (1) the proposed integrated hot, wet and steam pressure analysis method is feasible and has the versatility; (2) for the interface spall failure device dimensions and material selection is necessary to optimize the design. Due to the interaction between the parameters, the parameters of the device optimized combination is not unique, it is necessary to explore more optimal combination of design method which is suitable for this package devices. (3) BP neural network input factors using principal component analysis, to solve the problem of instability of the network fitting error and generalization ability (forecast); using principal component analysis and genetic algorithm improved BPNN model trained GA-BPNN model has good stability, and practical features. (4) optimal design method based on the improved BP neural network package devices a better solution to the design of the packaged devices with material and size with the problem. (5) In the parameter optimization process automation and user-oriented design software interface, the design system can be well packaged devices optimized design, the design process is convenient, simple and practical reference value. (6) Failure Cases optimization results show that the optimal design method is feasible. The results of this study toward engineering applications of simulation analysis and design of the microelectronics packaging device has a more important role.

Related Dissertations

CLC: > Industrial Technology > Radio electronics, telecommunications technology > Microelectronics, integrated circuit (IC) > General issues > Design
© 2012 www.DissertationTopic.Net Mobile