论文部分内容阅读
研究了神经网络-模糊推理协作系统在闸门综合自动化故障诊断中的应用,在该协作系统中,先根据模糊模型确定神经网络的结构、连接方式以及初始权值,从而构造出相应的神经网络.然后,将神经网络应用于闸门综合自动化故障诊断系统,根据系统的运行数据对神经网络进行学习训练,调整神经网络的权值来修正原始的模型,提高其准确性.经过学习训练后的神经网络模型又被转换成模糊模型,从而使分布存储在神经网络中的知识得到清晰的解释.该系统已应用于生产实践中,系统运行结果验证了该方法的有效性和可行性.
The application of neural network-fuzzy reasoning collaborative system in the gate automation fault diagnosis is studied. In this collaborative system, the structure, connection and initial weights of the neural network are first determined according to the fuzzy model to construct the corresponding neural network. Then, the neural network is applied to the sluice comprehensive automatic fault diagnosis system, the neural network is trained according to the operating data of the system, and the weight of the neural network is adjusted to correct the original model and improve its accuracy. After learning and training neural network The model is then converted into a fuzzy model, so that the knowledge stored in the neural network can be clearly explained. The system has been applied in production practice. The results of system operation verify the effectiveness and feasibility of the proposed method.