基于自主诊断重构技术的航天器故障检测系统设计

Design of Spacecraft Fault Detection System Based on Autonomous Diagnosis and Reconstruction Technology

摘要:

传统航天器故障检测系统姿态定位能力较差，导致不能突破阈值，准确实现检测，且传统系统不具备重构能力。为解决上述问题，基于自主诊断重构技术，提出了一种故障检测的新方法，集成故障检测器、数据采集器、滤波器等硬件设备。构建小波神经网络，结合航天器故障检测原理和重构问题集合神经网络算法和小波函数的优势构建小波神经网络，引入故障检测算法，实现航天器故障检测系统的自主诊断以及重构。实验结果表明，设计的基于自主诊断重构技术的航天器故障检测系统能够很好地从X、Y、Z三个轴进行检测，确定不同方位的航天器故障，在设定阈值后，提出的检测系统能够很好地分析阈值，实现残差突破，同时具备路线重构能力。

Abstract:

The traditional spacecraft fault detection system has poor attitude positioning ability, which makes it impossible to break through the threshold and accurately realize the detection, and the traditional system does not have the ability to reconstruct. In order to solve the above problems, based on the self-diagnosis and reconstruction technology, a new method of fault detection is proposed, which integrates hardware devices such as fault detectors, data collectors, filters and so on. Constructing wavelet neural network, combining the advantages of spacecraft fault detection principle and reconstruction problem set neural network algorithm and wavelet function, constructing wavelet neural network, introducing fault detection algorithm to realize autonomous diagnosis and reconstruction of spacecraft fault detection system. The experimental results show that the designed spacecraft fault detection system based on the self-diagnostic reconstruction technology can detect from the three axes of X, Y, and Z to determine the spacecraft faults in different azimuths. After setting the threshold, the proposed the detection system can analyze the threshold well, realize the breakthrough of the residual error, and has the ability of route reconstruction.