应用ＢＰ－ＡＲＴ混合神经网络提出了一种供推进系统状态监控实时使用的系统，其拓扑结构为： 第一层处理单元由ＢＰ神经网络组成， 每个ＢＰ网络代表一个相应的推进系统组件； 第二层处理单元为一个ＡＲＴ神经网络， 网络的每一个输出代表推进系统的一种“健康状态”， 据此可对其故障进行“诊断”。该混合结构充分发挥了两类网络的优点， 给出的具体应用实例也显示出在推进系统实时状态监控与故障诊断应用中的有效性 A BP-ART hybrid neural network is proposed to provide a real-time system for propulsion system condition monitoring. Its topology is as follows: the first layer processing unit is composed of BP neural network and each BP network represents a corresponding propulsion system component; The Layer 2 processing unit is an ART neural network, and each output of the network represents a “health state” of the propulsion system from which the fault can be “diagnosed.” The hybrid structure gives full play to the advantages of the two types of networks. The specific application examples given also show the effectiveness in the real-time condition monitoring and fault diagnosis of propulsion system