基于传感器网络的分布式故障检测与故障估计及其应用

In order to improve the autonomy, security and reliability of large-scale distributed systems, this project studies distributed fault diagnosis based on sensor networks, including distributed fault detection and fault estimation. Firstly, the model uncertainty, disturbances and sensor network topology are fully considered, and a distributed measurement model for fault diagnosis is built. Secondly, the influence of uncertainty and disturbance on distributed fault detection is analyzed. The dynamic threshold generation method of distributed autonomous fault detection based on zonotope is studied to improve fault detection efficiency and accuracy. Then considering packet loss and topology change of sensor networks, the distributed reliable fault estimation method is proposed based on sensor network to improve the accuracy and reliability of fault estimation. Finally, the proposed theoretical method is applied to the practical engineering system to verify the effectiveness of the method. This project aims to build a distributed fault diagnosis scheme based on sensor network, which enriches the theoretical content of distributed fault diagnosis and promotes the practical application of distributed fault diagnosis theory.

为了提高大型分布式系统的自主性、安全性和可靠性，本项目研究基于传感器网络的分布式故障诊断问题，具体包含分布式故障检测和故障估计。首先考虑被诊断系统的模型不确定性、干扰和传感器网络拓扑结构，建立面向故障诊断的分布式测量模型。其次，分析不确定性和干扰对分布式故障检测的影响，研究基于中心对称多胞体技术的分布式自主故障检测动态阈值生成方法，提高故障检测效率和准确率。继而考虑传感器网络通信丢包和拓扑结构变化的情形，研究基于传感器网络的分布式可靠故障估计方法，提升故障估计的精确性和可靠性。最后将所提的分布式故障诊断方法应用于工程实际系统，验证方法的有效性。本项目旨在构建一套基于传感器网络的分布式故障诊断方案，丰富分布式故障诊断理论内容，同时推动分布式故障诊断理论的工程应用。

本项目针对大型分布式系统的自主性、安全性和可靠性需求，开展了基于传感器网络的分布式故障诊断研究，具体包含分布式故障检测和故障估计。首先考虑被诊断系统的模型不确定性、干扰和传感器网络拓扑结构，建立了面向故障诊断的分布式测量模型。其次，分析不确定性和干扰对分布式故障检测的影响，研究了基于多项式展开的区间估计方法，提出了基于中心对称多胞体技术的分布式自主故障检测动态阈值生成方法，提高了故障检测效率和准确率。结合故障估计观测器设计方法，研究了基于传感器网络的分布式可靠故障估计方法，提升故障估计的精确性和可靠性，在此基础上，进一步探索了复杂系统智能容错控制方法的研究。最后将所提的分布式故障诊断方法应用于工程实际系统模型，验证方法的有效性。本项目构建了一套基于传感器网络的分布式故障诊断方案，为大型系统的分布式故障诊断工程应用奠定了基础。