| 引用本文: | 王远明,宫克,毛飞雄,杨明思,肖龙,李明辉.基于ACM及机器学习的腐蚀监测分析系统在重大基础设施上的应用[J].中国表面工程,2024,37(6):205~215 |
| WANG Yuanming,GONG Ke,MAO Feixiong,YANG Mingsi,XIAO Long,LI Minghui.Corrosion Monitoring and Analysis System for Critical Infrastructure based on ACM and Machine Learning[J].China Surface Engineering,2024,37(6):205~215 |
|
| |
|
|
| 本文已被：浏览 753次 下载 608次 |
码上扫一扫！ |
|
|
| 基于ACM及机器学习的腐蚀监测分析系统在重大基础设施上的应用 |
|
王远明1，2，宫克1，毛飞雄1，杨明思1，2，肖龙3，李明辉3
|
|
1.中国科学院宁波材料技术与工程研究所海洋关键材料重点实验室 宁波 315201 ；2.中国科学院大学材料科学与光电技术学院 北京 101408 ；3.宁波市杭州湾大桥发展有限公司 宁波 315300
|
|
| 摘要: |
| 腐蚀监测是确保基础设施安全的关键技术，通过实时跟踪材料的腐蚀状况来预防结构失效和环境事故，这项技术广泛应用于石油天然气、化工、船舶和桥梁等行业以维护资产的完整性和延长使用寿命。阐述一种基于电偶型腐蚀传感器（ACM） 原理的在线腐蚀监测技术在杭州湾跨海大桥上的应用，分析材料腐蚀监测技术应用现状，开发材料腐蚀与环境作用关系分析系统，并结合机器学习方法，包括随机森林、决策树、线性回归等方法，对监测数据进行挖掘，并对几种学习方法的训练效果与预测情况进行系统分析，得出随机森林训练效果最好，决定系数 R2达到 0.83，所开发系统采用无线传输方式，数据上传至云端，可同时对八种材料或涂层腐蚀进行监测，配备客户端软件进行数据处理分析，实现对各测试材料及涂层腐蚀失效时间的预测，并以此对重大设施关键点位的腐蚀状况进行实时监测。通过实时监测腐蚀情况能够及时发现潜在的安全隐患，从而采取措施，避免因腐蚀导致的重大安全事故，保障人员生命财产安全。基于对不同材料及涂层腐蚀特性的深入了解以及利用先进的数据分析手段可以制定出更为科学合理的维护计划，减少不必要的检查和维修成本，提升整体运营效率。引入最新的传感技术和机器学习算法为腐蚀监测领域带来了新的解决方案和技术思路，推动了相关行业的技术进步与发展，而且有效控制和减缓腐蚀过程，有助于降低由于设备故障而造成的环境污染风险，符合可持续发展的要求。对腐蚀程度的精准评估可以帮助企业合理安排资源分配，避免过早更换仍具有良好性能的部件，从而节约大量资金投入。该研究不仅具有实际应用价值，也为腐蚀机理的研究提供了宝贵的试验数据支持，丰富了关于材料耐蚀性的理论体系，对于进一步探索新型防腐材料和技术具有重要意义。 |
| 关键词: 材料腐蚀监测 环境参数 涂层 机器学习 |
| DOI：10.11933/j.issn.1007-9289.20240111001 |
| 分类号:TG172 |
| 基金项目:国家重点研发计划（2022YFC3103803）；中国科学院国际合作项目（174433KYSB20200006）；宁波市重点科技项目（2021Z079） |
|
| Corrosion Monitoring and Analysis System for Critical Infrastructure based on ACM and Machine Learning |
|
WANG Yuanming1，2，GONG Ke1，MAO Feixiong1，YANG Mingsi1，2，XIAO Long3，LI Minghui3
|
|
1.Key Laboratory of Advanced Marine Materials, Ningbo Institute of Materials Technology and Engineering,Chinese Academy of Sciences, Ningbo 315201 , China ；2. College of Materials Science and Opto-Electronic Technology, University of Chinese Academy of Sciences,Beijing 101408 , China ；3. Ningbo Hangzhou Bay Bridge Development Co., Ltd., Ningbo 315300 , China
|
| Abstract: |
| Corrosion monitoring is a crucial technology that ensures the safety and longevity of infrastructure by providing real-time tracking of the corrosion conditions of materials, helping prevent structural failures and major accidents. This technology has been extensively applied in various industries, including oil and gas, chemical engineering, shipping, and bridge construction, with the primary aims of ensuring operational safety and prolongating the service life of the equipment. This study focuses on the application of an advanced online corrosion-monitoring technology based on the principle of galvanic-type corrosion sensors, known as Advanced Corrosion Monitoring (ACM), specifically in the context of cross-sea bridges, which are subject to harsh environmental conditions such as saltwater, high humidity, and fluctuating temperatures, making them particularly susceptible to corrosion. The ACM system was designed to continuously monitor the corrosion processes of the various materials and coatings used in these structures, thereby enabling timely interventions and maintenance. The current landscape of material corrosion monitoring and detection technologies is diverse and rapidly evolving, complementing traditional methods such as visual inspection and periodic sampling and, in some cases, replacing more advanced and continuous monitoring systems that leverage a range of sensors and data collection techniques to provide real-time, accurate, and comprehensive information about the corrosion status of materials. Ultrasonic testing, eddy current testing, and electrochemical impedance spectroscopy are typically used to detect and quantify corrosion. To better understand the relationship between material corrosion and environmental factors, an analysis system was developed, integrating multiple environmental parameters and collected data on temperature, humidity, salinity, pH, and other relevant environmental variables. By correlating these environmental parameters with the observed corrosion rates, the system identifies key factors influencing the corrosion process and predicts future corrosion trends. For instance, high humidity and salt content in air are known to accelerate the corrosion of steel, whereas low temperatures can slow the corrosion rate. Machine learning methods, including random forests, decision trees, and linear regression, have been incorporated to enhance the predictive capabilities of corrosion-monitoring systems. These algorithms were trained on historical data to uncover close relationships between environmental data and material corrosion using random forests, which showed the best performance in the study, achieving a determination coefficient R2 of 0.83 and demonstrating a high level of accuracy in predicting corrosion rates. While decision trees provide a clear and interpretable model, allowing us to understand the decision-making process behind the predictions, they may not achieve the same level of accuracy as random forests. However, they offer valuable insights into the key factors influencing corrosion. Linear regression, although simpler, provides a useful baseline for comparison, elucidating the linear relationships between the environmental parameters and corrosion rates. The ACM-based system adopts a wireless data transmission mode and uploads real-time data to the cloud, which allows for centralized data storage, processing, and analysis. It is equipped with client software that provides a user-friendly interface for data visualization and analysis. Key features include real-time monitoring of corrosion conditions at critical points in the infrastructure, enabling immediate detection of anomalies; advanced data processing and analysis tools to extract meaningful insights from the collected data; predictive models that estimate the remaining service life of materials and coatings, helping in planning maintenance activities; automated alerts and notifications to inform stakeholders of potential issues and recommended actions; and customizable reports that can be generated to provide detailed insights into the corrosion status and trends over time. Case studies on cross-sea bridges in various regions illustrated the effectiveness of ACM systems. For example, a bridge in a coastal area with high salinity and humidity levels was monitored using the ACM system, and the data collected over a period of two years showed a significant correlation between the environmental parameters and corrosion rates of the bridge’s steel components. By implementing the recommendations provided by the system, bridge maintenance teams can reduce the corrosion rate and extend the service life of the structure. Another case study involving a bridge in a region with varying climatic conditions detected an unexpected increase in corrosion rates during a specific season; further analysis revealed that this was due to a combination of high humidity and increased pollution levels. The maintenance team took immediate action, applied protective coatings, and implemented additional measures to mitigate the corrosion. The ACM system provides several benefits. It not only enhances the safety and reliability of infrastructure but also reduces maintenance costs by enabling proactive and targeted interventions. The ability of the system to predict the remaining service life of the materials and coatings enables better planning and resource allocation. In the future, ACM systems can be further enhanced by integrating more advanced sensors, such as fiber-optic sensors, and by incorporating more sophisticated machine learning algorithms. In addition, the system can be extended to other types of infrastructure, such as pipelines, offshore platforms, and industrial facilities, to provide a comprehensive solution for corrosion management. In conclusion, the application of advanced corrosion-monitoring technologies, such as the ACM system, is essential for ensuring the safety and longevity of infrastructure, particularly in challenging environments such as those encountered in cross-sea bridges. By integrating machine learning methods and real-time data analysis, these systems provide valuable insights and predictive capabilities, enabling proactive maintenance and preventing major accidents. As technology evolves, it promises to further enhance the reliability and efficiency of infrastructure management. Keywords: corrosion monitoring, environmental parameters, coatings, machine learning, random forest, decision tree, linear regression, cross-sea bridge, infrastructure safety, service life extension material corrosion monitoring; environmental parameters; coatings; machine learning |
| Key words: material corrosion monitoring environmental parameters coatings machine learning |
|
|
|
|
