【摘要】 控制系统的不确定故障诊断是目前尚未解决的问题。传统的神经网络方法和模糊推理方法为解决这一类故障诊断问题提出了一些算法。然而上述两种方法难以提高不确定故障诊断的性能。为此本文结合模糊理论的推理能力和神经网络学习的能力提出模糊神经网络故障检测方法。该算法同时具备模糊理论的处理不确定和不准确信息的能力和神经网络的自学习能力。本文结果应用到某运载火箭控制系统的故障诊断,仿真结果表明,本算法有效,较好地解决了控制系统不确定故障诊断问题。更多还原

【Abstract】 The undecided fault in control system is a problem that has not been solved. The traditional methods using neural network and fuzzy reasoning provide some algorithms to this problem. But it is difficult to improve the performance of fault diagnosis in the field above. In this paper, a method based on both fuzzy theory and neural network is put forward. This method combines the capability of fuzzy reasoning in settling uncertain and imprecise information and the capability of neural networks in learning from examples. The simulation result and the application of this method in the launch vehicle control system verify the efficiency of the algorithm.更多还原