基于稀疏数据规约的CMAC大气腐蚀数据补偿方法  

Missing data compensation method in atmospheric corrosion data with CMAC based on sparse data regular

作　　者：李志平[1] 付冬梅[1] 穆志纯[1] Li Zhiping;Fu Dongmei;Mu Zhichun(School of Automation & Electrical Engineering, University of Science & Technology Beijing, Beijing 100083 , China)

机构地区：[1]北京科技大学自动化学院,北京100083

出　　处：《计算机应用研究》2016年第9期2645-2647,共3页Application Research of Computers

基　　金：国家重点基础研究发展"973"计划资助项目(2014CB643300);国家材料环境腐蚀平台资助项目

摘　　要：材料腐蚀带来巨大的损失。对于大部分地区来说,大气腐蚀等级是未知的。如何准确地补偿缺失的大气腐蚀等级数据成为函待解决的问题。分别针对大气腐蚀等级的两个关键因素氯离子浓度、二氧化硫浓度单独缺失的情况进行数据建模补偿。针对中国地区腐蚀等级以及相关环境参数数据稀疏分布的特性,提出了一种基于稀疏数据规约的CMAC大气腐蚀数据补偿方法。同时,针对二氧化硫浓度缺失的情况,根据现有的有效数据,提出了二氧化硫浓度的经验公式。结果表明,氯离子浓度单独缺失时预测准确率为86.5%,二氧化硫浓度单独缺失时预测准确率为82.6%。该算法提高了大气腐蚀等级数据补偿准确率,为材料选择提供了重要依据。Material corrosion had brought huge loss. The atmospheric corrosion levels were unknown for most of the regions.How to compensate accurately for atmospheric corrosion data was a problem which needed to be solved. Two key factors which determined the atmospheric corrosion level, chlorine ionconcentration and sulfur dioxideconcentration, usually did not have da ta.This paper solved the problem by imputing missing data. Atmospheric corrosion level data and relevant environmental parametersdata in China were both sparse distribution. According to the characteristic, this paper put forward a missing data compensation method in atmospheric corrosion data with CMAC based on sparse data regular. For the situation of imputing sulfurdioxide concentration data, this paper put an empirical formula of sulfur dioxide concentration. The results showed that the prediction accuracy was 86.5% when the chloride ionconcentration was absent, and the prediction accuracy was 82. 6% when the sulfur dioxide concentrationwas absent. This algorithm can improve the prediction accuracy of atmospheric corrosion levels data, and provides an important basis for the selection of material.

关 键 词：缺失数据 大气腐蚀等级 小脑模型 稀疏数据规约 二氧化硫浓度经验公式 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]