论文部分内容阅读
在轨航天器故障检测与诊断问题需要面对模型的非线性,而且要求尽量提高其检测的精度,为此设计了基于径向基函数(RBF)神经网络的动量轮非线性故障检测与诊断(FDD)方案。首先应用RBF补偿建模误差,提高检测精度,并选择李雅普诺夫函数证明其收敛性;然后应用非线性观测器来产生故障残差,给出了阈值以及故障检测的时间;应用RBF网络对故障信号进行重构,并据此设计了带学习能力的FDD策略。再次建立了详细的动量轮模型,通过不同条件下的仿真研究分别验证残差的阈值特性、时间特性以及RBF的重构能力,仿真结果表明了算法的有效性。
In spacecraft fault detection and diagnosis, it is necessary to face the nonlinearity of the model, and the precision of the detection is required to be as high as possible. For this reason, a nonlinear RBF neural network based fault diagnosis and diagnosis FDD) program. At first, RBF is used to compensate the modeling errors to improve the detection accuracy. The Lyapunov function is used to prove its convergence. Then, the nonlinear observer is used to generate the residual error, and the threshold and fault detection time are given. Signal is reconstructed, and accordingly FDD strategy with learning ability is designed. A detailed model of momentum wheel is established again. The threshold characteristics, time characteristics and RBF reconstruction capability of the residual are validated by simulation under different conditions. The simulation results show the effectiveness of the algorithm.