针对水轮机状态检测系统检测量多,实时性强等特点,本文提出了一种基于信息融合技术的神经网络专家系统,它通过基于信息融合的故障分类子网络技术克服了单层神经网络难以添加新的故障类型的弱点,综合了专家系统和人工神经网络的各自的优点,充分利用专家系统的推理判断能力和人工神经网络的学习和容错能力,从而使整个状态检测系统具有更好的性能。 Aiming at the characteristics of turbines state detection system such as large amount of detection and strong real-time performance, this paper proposes a neural network expert system based on information fusion technology, which overcomes the difficulty of adding single layer neural network The fault types of the weaknesses, the combination of expert systems and artificial neural network of their respective advantages, and make full use of expert system of reasoning and judgment and artificial neural network learning and fault tolerance, so that the whole state detection system has better performance.