【作者】 宋小安；

【导师】 李志华；

【作者基本信息】 河海大学 ， 控制理论与控制工程， 2005， 硕士

【摘要】 随着电子工业的迅速发展，模拟电路故障诊断技术的重要性越来越明显，它对于电子设备或系统的正常运行和可靠性设计均具有重要的意义。在传统的诊断技术和理论方法的基础上，本课题以雷达电源为研究对象，就专家系统和BP神经网络方法应用于模拟电路故障诊断作了深入的研究。 详细阐述了雷达电源故障诊断专家系统的构造过程，包括知识的获取、表达和存储以及推理机的推理方式和控制策略等。采用Visual C++6.0作为开发工具，专家系统诊断使用友好的人机交互式界面，产生式规则表示法，关系型数据库的知识存储方式，树状的诊断流程作为诊断模型，基于不确定性和数据驱动的正向推理，深度优先和启发式相结合的搜索策略，同时还具有完备的知识获取和解释机制。 BP神经网络具有联想记忆功能、容错性、鲁棒性以及很好的非线性映射能力，本文具体描述了神经网络样本的获取、网络结构的确定和训练。通过运用粗糙集理论中的属性约简来处理样本，从而达到简化网络结构的目的；通过运用电路仿真软件改善训练样本，最终提高网络的泛化能力。 通过实践证明，以上两种方法均行之有效，对于模拟电路的故障诊断具有一定的应用价值和发展潜力。更多还原

【Abstract】 With the rapid development of the Electronic Industry, the importance of the analog circuit fault diagnosis is more and more obvious, it has important signification for working orderly and dependability design of electronic equipment or system. On the base of traditional diagnosis and theory, Take the radar power as studying object, this paper studies the application of Expert System and BP neural network in the analog circuit fault diagnosis.Particularly expatiates the constitution of the radar power Expert System, including ingathering, expression and memory of knowledge, reasoning mode and controlling strategy of reasoning mechanism. Adopts Visual C++ 6.0 as developing tools, Expert System diagnosis uses friendly man-computer dialog platform, producing method, relationship database memory mode, fault tree diagnosis flow as diagnosis model, forward reasoning bases on uncertainty and data drive, search combines depth first and heuristic strategy, at the same time, it has self-contained mechanism for knowledge ingathering and explain.BP neural network has the robustness, associated memory and nonlinear mapping, this paper embodies ingathering of samples, confirmation of framework and training of network. Adopts attribute reduction of Rough Set to deal with training samples, predigest samples to predigest framework; Improve samples with circuit emulator to improve recognition ability of network.Practice has proved that the two methods mentioned above are all effective, they have great applying value and developing potential in analog circuit fault diagnosis.更多还原