摘要
核动力装置监控系统中,由于数据采集和传输过程中受到外部干扰或由于设备自身问题,导致在运行参数监测数据中存在数据缺失现象,这将很大程度影响对运行数据分析结果的正确性。为解决此问题,首先基于核动力装置某一特定的运行状态对应的各组运行参数之间有着极高相似性的特点,开展了基于热卡填充法的核动力装置运行数据缺失值填充方法研究,并利用运行参数与系统状态对应关系,实现了缺失值的估计;然后,设计了系统状态表达方法、状态相似性测量方法以及缺失数据填充算法总体流程;最后,以模拟机运行数据为例,构建了测试集进行测试。测试结果证明了所设计算法的可行性与准确性。
Data missing is one of the most serious data quality problems in nuclear power plant.In or-der to solve the problem,the filling method for missing data of operating parameters of nuclear power plant based on hot-decking(HD)algorithm was studied,since there was a very high similarity among operating parameters corresponding to a specific operating state of nuclear power plant.And it made use of the corresponding relationship between operation parameters and system state to estimate the missing data.The system state expression method,the expression method for parameter vector simi-larity measurement and the overall flow of data filling algorithm were designed.Finally,taking the simulator operation data as an example,the test set was constructed for testing whereby the feasibility and accuracy of the proposed algorithm were proved.
作者
毛伟
余刃
王天舒
殷少轩
MAO Wei;YU Ren;WANG Tianshu;YIN Shaoxuan(College of Nuclear Science and Technology,Naval Univ.of Engineering,Wuhan 430033,China;Security Evaluation and Support Office of Naval Staff,Beijing 100166,China)
出处
《海军工程大学学报》
CAS
北大核心
2023年第3期8-14,共7页
Journal of Naval University of Engineering
基金
国家重点实验室基金资助项目(LRSDT2017302)。
关键词
核动力装置
缺失值填充
马氏距离
热卡填充法
nuclear power plant
missing data filling
Mahalanobis distance
hot-decking algorithm
