摘要
基于人工经验的核电厂变电站电力调度操作票校验方法具有主观性强、校验效率低、可靠性不高等问题,为解决上述问题,提出了一种基于卷积神经网络(CNN)和双向长短时记忆网络(BiLSTM)的操作票自动校验方法。该方法在对操作票文本进行向量化的基础上,利用CNN-BiLTSM模型实现操作票文本的深度信息挖掘和校验的自动化。实验结果表明,相比单一模型,CNN-BiLSTM模型的校验精度更高,校验评估综合指标可达95.67%,具有一定的优势。
The verification method of nuclear power plant power dispatching operation ticket based on manual experience has some problems,such as strong subjectivity,low efficiency and low reliability.To solve these problems,an automatic verification method of operation ticket based on Convolutional Neural Network(CNN)and Bidirectional Long Short Term Memory(BiLSTM)was proposed.Based on vectorization of the operation ticket text,this method used the CNN-BiLSTM model to realize the automation of deep information mining and verification of operation ticket text.The experimental results show that compared with the single model,the CNN-BiLSTM model has higher calibration accuracy,and the comprehensive index of calibration evaluation can reach 95.67%,which has certain advantages.
作者
吴奇珂
程培军
钱韦廷
姜浩宇
胡佳
WU Qike;CHENG Peijun;QIAN Weiting;JIANG Haoyu;HU Jia(Guangdong Power Grid Co.,Ltd.Guangzhou Power Supply Bureau Power Dispatching and Control Center,Guangzhou Guangdong,510620,China)
出处
《核电子学与探测技术》
CAS
北大核心
2024年第2期316-322,共7页
Nuclear Electronics & Detection Technology
基金
南方电网广东广州供电局科技项目(030108KK52222002),项目名称:基于智能语义识别和高级安全防误的启动方案程序化操作技术研究。
关键词
卷积神经网络
操作票
自动校验
双向长短时记忆网络
Convolutional Neural Network
operation ticket
automatic verification
bidirectional long short term memory
