On-line Monitoring Method of Large-scale Weapon Equipment Based on Multilayer Competition Neural Network

The large-scale weapon equipment is often a complex system, which applies many techniques such as mechanical technique, electronics, hydraulic and so on. Whole weapon system will lose combat capability, if the large-scale weapon equipment is fault. It is necessary to realize real-time monitoring and automatic recognition of working state, to rapidly locate fault´s components for the large-scale weapon equipment. For this reason, this method that adopts the competition network to realize on-line monitoring is put forward. When the fault has been found, the fault diagnosis is completed by calling the corresponding neural network. By this method, not only the neural network´s scale is reduced, but also the design´s complexity of diagnosis system is simplified and computing time is reduced too. This method has an important significance to realizing on-line monitoring and diagnosis of the large-scale weapon equipment.