基于脉冲神经膜系统的智能输电网故障诊断模型及其自动设计方法研究

Spiking neural P systems is a new type of high performance distributed parallel neural-like computing models inspired by the spike processing mechanism of brain neurons and show great potential in solving the fault diagnosis of smart transmission grids. Lacking fault diagnosis models and their automatic design methods is one of the most important problems for the application of spiking neural P systems to smart transmission grid fault diagnosis. This project will systematically and deeply investigate the design method of fault diagnosis models based on spiking neural P systems for smart transmission grids from three aspects: fault diagnosis mechanism, fault diagnosis models and their automatic design methods. This project focuses on exploring the intrinsic association between the working mechanism of spiking neural P systems and the fault diagnosis mechanism of smart transmission grids. Through deeply investigating the mechanism and models of smart transmission grid fault diagnosis based on spiking neural P systems, constructing the fault information base and the P system ingredient base, designing effective modeling criterions and methods, and the testing, analysis and evaluation approaches for the built models, this project accomplishes the objective that the automatic design method of fault information-driven spiking neural P system fault diagnosis models is presented. This project has important academic value and engineering application significance because it will not only introduce novel and high performance intelligent diagnosis models and their automatic design methods to smart transmission grid fault diagnosis , but will also offer a new thought for theoretical and application research of membrane computing.

脉冲神经膜系统是受大脑神经元处理脉冲机制启发，而建立的新型高性能分布式并行计算模型，在智能输电网故障诊断领域具有巨大应用前景。缺乏故障诊断模型及其自动设计方法，是其在该领域应用中面临的重要问题之一。本项目拟从诊断机理、诊断模型和模型自动设计三方面，系统深入地研究脉冲神经膜系统智能输电网故障诊断模型的构建方法。着力探索脉冲神经膜系统工作机制与智能输电网故障诊断机理之间的内在关联关系。通过深入研究基于脉冲神经膜系统的智能输电网故障诊断机理和模型，构建故障信息库和膜系统要素库，设计有效建模准则和方法，并对所构建模型进行测试、分析与评价，实现提出故障信息驱动的脉冲神经膜系统智能输电网故障诊断模型自动设计方法这一目标。研究成果既将为智能输电网故障诊断领域引入新型高性能智能诊断模型及其自动设计方法，也为膜计算的理论和实际应用研究提供新思路，具有重要理论学术价值和工程应用意义。

脉冲神经膜系统在输电网故障诊断领域具有巨大应用潜力，但是缺乏故障诊断模型及其构建方法是面临的重要难题。本项目致力于研究基于脉冲神经膜系统的电网故障诊断模型：(1)通过研究电网故障路径传播规律与脆弱性评估方法，获得了生物神经系统工作机制与输电网故障诊断机理之间的内在关联关系，建立了基于脉冲神经膜系统的故障诊断理论，为诊断模型的设计和构建方法的提出奠定了基础；(2)主要针对故障警报信息不确定性和电网拓扑变化自适应性等问题，针对性地提出了考虑生物凋亡机制的自更新脉冲神经膜系统、二进制脉冲神经膜系统、权值修正脉冲神经膜系统等系列性诊断模型，设计了对应的故障信息处理算法，并提出了相应的模型构建方法。本项目的研究成果为输电网故障诊断领域引入了新型高性能智能诊断模型及其构建方法，同时也可为其他领域的故障诊断提供借鉴和参考。