摘要
针对稳压器泄漏难以监测的问题,将集成神经网络(INN)与模糊逻辑融合(FLF)方法相结合,研究了稳压器泄漏监测方法。在该方法中,利用RBF神经网络(RBF-NN)建立稳压器泄漏诊断模型;采用两个Elman神经网络(Elman-NN)分别建立稳压器参数预测模型与稳压器泄漏诊断模型;应用模糊逻辑融合方法对RBF-NN与Elman-NN诊断的结果进行融合,并将其作为稳压器泄漏最终监测结果。为验证该方法的可行性,利用压水堆核动力装置模拟器对其监测效果进行了验证。结果表明,与单神经网络诊断方法相比,所提出的监测方法具有更高的可靠性;与其他泄漏监测方法相比,该方法简便、易行。
A new leak monitoring method based on integration neural networks(INN)and the fuzzy logic fusion(FLF)was proposed to solve the problem of pressurizer leak monitoring.In this approach,apressurizer leaking diagnosis model was established by a radial basis function neural network(RBF-NN).Two Elman neural networks(ElmanNN)were used to establish pressurizer parameters prediction model and pressurizer leak diagnosis model respectively.Then,the fuzzy logical method was used to fuse the diagnosed results of RBF-NN and Elman-NN.The fusion results were the final monitoring results.The nuclear power simulator was used to test the feasibility of the proposed method.The results show that compared with the diagnosis method based on single neural network,the proposed method is simple and reliable.
出处
《原子能科学技术》
EI
CAS
CSCD
北大核心
2014年第S1期474-479,共6页
Atomic Energy Science and Technology
关键词
稳压器
集成神经网络
模糊逻辑融合
泄漏监测
pressurizer
integrated neural networks
fuzzy logic fusion
leak monitoring
