论文部分内容阅读
凝汽器工作状况好坏直接影响凝汽式汽轮发电机组运行的安全性和经济性。分别采用自适应梯度递减学习算法的BP网络和Elman网络用于凝汽器的故障诊断。将测试样本输入已经学习好的BP网络和El-man网络中,测试结果表明这两种网络正确诊断故障的概率高达90%以上,但采用Elman网络的学习速度要比BP网络快,适合实时操作的场合。
Condenser working conditions directly affect the performance of condensing steam turbine running safety and economy. BP network and Elman network with adaptive gradient descending learning algorithm are respectively used for fault diagnosis of condenser. The test samples are input into the already learned BP network and El-man network. The test results show that the probability of correctly diagnosing the fault is higher than 90% for both networks. However, Elman network can learn faster than BP network and is suitable for real-time operation The occasion