论文部分内容阅读
针对电子系统缓变故障的预测问题,提出一种自适应相关向量机(RVM,Relevance Vector Machine)方法.首先,对反映电子系统性能的参数序列进行相空间重构,建立RVM的输入输出对应关系;然后,将嵌入维数和核函数参数作为人工鱼位置,取留一交叉验证(LOOCV,Leave-One-Out Cross-Validation)误差的相反数作为目标函数,利用人工鱼群算法(AFSA,Artificial Fish Swarm Algorithm)实现方法参数的自适应优化选择;最后,通过雷达发射机高压电源与多注速调管的故障预测实验验证了方法的性能.实验结果表明:该方法在预测精度和预测可靠性方面优于现有方法.
In order to solve the problem of gradual change fault in electronic system, an adaptive vector correlation machine (RVM) method is proposed.Firstly, phase space reconstruction of parameter sequence reflecting the performance of electronic system is carried out to establish the relationship between input and output of RVM Then, the embedding dimension and kernel function parameters are regarded as the artificial fish position, taking the opposite numbers of the LOOCV (Leave-One-Out Cross-Validation) error as the objective function. The artificial fish swarm algorithm (AFSA, Artificial Fish Swarm Algorithm) to achieve the optimization of the method parameters.Finally, the performance of the method is verified by the fault prediction of the radar transmitter high voltage power supply and the multi-injection klystrons.The experimental results show that the proposed method can be used in prediction accuracy and prediction reliability In terms of better than the existing methods.