针对单个神经网络的不足,以信息融合技术为基础,建立了集成神经网络故障诊断系统。该系统采用子神经网络从不同侧面对设备进行初步诊断,然后将各子神经网络进行决策融合。通过对汽轮发电机组的应用仿真,说明该系统充分利用了各种信息,提高了确诊率。 关键词:故障诊断;汽轮机组;集成神经网络;信息融合 Abstract:In view of the disadvantage of single neural network, this paper establishes an integrated neural network fault diagnosis system based on information fusion technology. This system uses the sub-neural network to carry on the preliminary diagnosis from the different sides to the equipment, and then carries on the policy-making fusion of each sub-neural network. The simulation of steam turbine generator sets’ fault diagnosis demonstrates that the system takes full advantages of diversified information and improves diagnosis precision. Key words: fault diagnosis; steam turbine-generator set; integrated neutral network; information fusion