With the constant development of avionics system, complexity and crucial features of the system are unceasingly enhanced. So more emphasis has been laid on the real-time online fault diagnosis of the system. Considering the characteristic of non-linear, complexity, strong interference and diversity showed in fault diagnosis in electronic system, the paper introduces a method of using support vector machine to diagnose the fault in avionics system; Meanwhile the paper utilizes the particle swarm optimization (PSO) algorithm to achieve the parameter optimization of support vector machine, which could improve the efficiency of choosing parameters and avoid the deficiency of choosing parameters artificially. The simulation result shows that this method merges the multipoint test information of the avionics system with simple structure and high efficiency. The fault detection rate reaches 97. 5% and the average fault recognition rate reaches 96. 9%. The method is suitable for online intelligent fault diagnosis of data fusion avionics system.%随着航空电子系统的不断发展,复杂性和关键性不断增强,其故障的实时在线诊断越来越受到重视；针对电子系统在故障诊断中表现出的非线性、复杂性、强干扰性和多样性的特点,提出采用支持向量机进行航空电子系统的故障诊断；同时,采用粒子群优化(PSO)算法实现支持向量机的参数寻优,以提高其参数选择的效率,避免人为选择参数的不足；仿真实验表明,该方法融合航空电子系统的多点测试信息,结构简单时效性高,故障检测正确率达到97.5％,平均故障识别正确率达到96.9％,适用于信息融合型的航空电子系统在线智能故障诊断.
展开▼
