Please wait a minute...
中国腐蚀与防护学报    DOI: 10.11902/1005.4537.2026.184
  本期目录 | 过刊浏览 |
基于随机森林的亚热带海洋环大气境腐蚀数据挖掘研究
何致远, 杨劲怀, 董衍旭, 乐艺泽, 何峰彬, 杨启帆, 夏晓健, 严康骅, 杨小佳

1国网福建省电力有限公司信息通信分公司,福建,泉州, 350001

2国网福建省电力有限公司电力科学研究院,福建,福州, 350007

3北京科技大学新材料技术研究院,北京,100083

Research on Subtropical Marine Atmospheric Corrosion Data Mining Based on Random Forest
Yang Xiao-Jia

¹ Information and Communication Branch, State Grid Fujian Electric Power Co., Ltd., Quanzhou, Fujian, 350001, China

² Electric Power Research Institute, State Grid Fujian Electric Power Co., Ltd., Fuzhou, Fujian, 350007, China

3 Institute for Advanced Materials and Technology, University of Science and Technology Beijing, Beijing, 100083, China

引用本文:

何致远, 杨劲怀, 董衍旭, 乐艺泽, 何峰彬, 杨启帆, 夏晓健, 严康骅, 杨小佳. 基于随机森林的亚热带海洋环大气境腐蚀数据挖掘研究[J]. 中国腐蚀与防护学报, 10.11902/1005.4537.2026.184.
, , , , , , , , Xiao-Jia Yang. Research on Subtropical Marine Atmospheric Corrosion Data Mining Based on Random Forest[J]. Journal of Chinese Society for Corrosion and protection, 0, (): 0-.

全文: PDF(1580 KB)  
摘要: 随着海上风电的快速发展,其基础设施在高温、高湿、高盐雾的亚热带海洋环境下面临严峻的腐蚀挑战。为探究腐蚀规律并实现智能预测,本研究在福建省泉州石圳海上风电场,利用智能大气腐蚀监测仪,对碳钢材料进行了为期一年的腐蚀电流及多种环境因子(温度、湿度、污染物等)的同步在线监测,构建了包含约50万条数据的大规模数据集。通过对数据进行清洗与融合,本研究首次将随机森林机器学习算法应用于该区域海洋大气腐蚀分析,建立了碳钢腐蚀预测模型,并利用SHAP等方法进行了特征重要性排序与模型可解释性分析。结果表明,湿度、氯离子沉降速率和温度是影响碳钢腐蚀的最关键环境因子;模型预测结果与实测值吻合良好,决定系数R²达0.89。该研究为理解亚热带海洋环境下碳钢的腐蚀机理、实现腐蚀速率预测以及海上风电设备的差异化选材与防护策略制定提供了数据支撑与智能方法。
关键词 : 海上风电,  腐蚀预测,  随机森林    
Abstract:With the rapid development of offshore wind power, its infrastructure faces severe corrosion challenges in the subtropical marine environment characterized by high temperature, high humidity, and high salt spray. To investigate the corrosion patterns and achieve intelligent prediction, this study conducted synchronized online monitoring of corrosion current for carbon steel materials and various environmental factors (e.g., temperature, humidity, pollutants) over one year at two offshore wind farms in Shizhen Quanzhou, Fujian Province, using intelligent atmospheric corrosion monitors. A large-scale dataset containing approximately 500,000 data entries was constructed. Through data cleaning and fusion, this research applied the Random Forest machine learning algorithm for the first time to the analysis of marine atmospheric corrosion in this region, establishing a carbon steel corrosion prediction model. Feature importance ranking and model interpretability analysis were performed using methods like SHAP. The results indicate that humidity, chloride deposition rate, and temperature are the most critical environmental factors affecting carbon steel corrosion. The model's predictions show good agreement with measured values, achieving a coefficient of determination (R²) of 0.89. This study provides data support and an intelligent methodology for understanding the corrosion mechanism of carbon steel in subtropical marine environments, predicting corrosion rates, and formulating differentiated material selection and protection strategies for offshore wind power equipment.
Key words: offshore wind power    corrosion prediction    Random Forest
收稿日期: 2026-06-22     
[1] 赵梓恒, 李波, 杨臻, 钱旺, 于浩, 张昊, 易盼, 陈俊航, 尹程辉, 肖葵. 海南滨海大气环境热浸镀锌钢室内外试验相关性分析[J]. 中国腐蚀与防护学报, 2026, 46(4): 1129-1138.
[2] 张明, 龚宁, 张博, 霍宏博, 李浩男, 钟显康. 基于腐蚀预测和强度计算的油管服役寿命评估[J]. 中国腐蚀与防护学报, 2024, 44(3): 781-788.
[3] 朱烨森, 蔡锟, 胡葆文, 夏云秋, 胡涛勇, 黄一. 海底管道CO2 腐蚀特性及预测模型研究进展[J]. 中国腐蚀与防护学报, 2023, 43(6): 1225-1236.
[4] 王莎莎, 马帅杰, 车琨, 杜艳霞. 机器学习在自然环境腐蚀评估与预测领域的应用现状[J]. 中国腐蚀与防护学报, 2023, 43(3): 441-451.
[5] 王炳钦, 张晓莲, 雍兴跃, 周欢, 高新华. 舰船海水管系中紫铜/钢制管道耦接后电偶腐蚀的数值模拟研究[J]. 中国腐蚀与防护学报, 2022, 42(2): 200-210.
[6] 刘静,李晓禄,朱崇伟,张涛,曾冠鑫,孟国哲,邵亚薇. 利用人工神经网络技术预测气田环境下316L不锈钢临界点蚀温度[J]. 中国腐蚀与防护学报, 2016, 36(3): 205-211.
[7] 韩夏冰 高志明 党丽华 王 迎 毕慧超. Q235钢在模拟大气环境中
早期腐蚀图像小波包分析
[J]. 中国腐蚀与防护学报, 2013, 33(3): 211-215.
[8] 谭晓明 . 基于人工神经网络的铝合金腐蚀预测及其精度分析[J]. 中国腐蚀与防护学报, 2004, 24(4): 218-221 .
[9] 崔大为; 林乐耘; 赵月红 . 应用人工神经网络预测有色金属海水腐蚀的长期行为[J]. 中国腐蚀与防护学报, 2004, 24(1): 29-32 .
[10] 蔡建平;柯伟. 应用人工神经网络预测碳钢、低合金钢的大气腐蚀[J]. 中国腐蚀与防护学报, 1997, 17(4): 303-306.