In any kind of industry where water meets steel, corrosion is an underestimated threat - and a very expensive one too: corrosion costs for the European region alone amount to 500 billion euro, of which up to 1/3 can be saved according to a NACE IMPACT-study. In the past, corrosion management was typically focussed on periodic inspections of the entire asset or fleet, and hopefully fix things before they lead to downtime, safety or environmental hazards. To realise the savings suggested in the NACE study, the SOCORRO project aims to provide a framework for data driven and more efficient corrosion management. In this blog, I want to give you an eagle-eye view of the approach being developed in the SOCORRO project. Keep following our activities and learn more in future blogs!
Blog written by Jeroen Tacq, Sirris
A change in mindset on Corrosion Management
In recent years, the mindset on how to manage corrosion effectively is quickly changing. Corrosion management can be made more efficient using data, allowing to target (parts of) assets having the highest risk of failure due to corrosion. New Non-Destructive Testing (NDT) and Non-Intrusive Inspection (NII) approaches for screening of assets are therefore being developed, often involving the use of drones, ROVs (Remote Operated Vehicles) and innovative inspection techniques. However, these techniques still often are limited to spot inspections and don’t give online, instant information.
To overcome those limitations, the SOCORRO project explores the possibility to use data about the assets’ environment to predict the overall corrosion risk. Through a combination of collecting sensor data (either fixed in place or on a moving platform) and Artificial Intelligence (AI), the SOCORRO project aims to establish a decision support tool for data driven corrosion management. These tools should not be seen as a way to replace engineers, rational thinking and decision makers, but rather as tools providing data for the decision-making process.
Defining Corrosion Risk
The exact rate at which corrosion occurs is notoriously difficult to predict and can be extremely location dependent. On the one hand, that means the choice of sensor position is critical, on the other hand, it means that an exact prediction of corrosion rates is likely to deviate from reality and lead to erroneous decision. Therefore, within the SOCORRO approach, we calculate a corrosion risk, based on measurements at certain points, without explicitly having corrosion rate as an output. A risk of ‘1’ can be thought of as the risk level that leads to fully consuming the corrosion allowance within the expected lifetime.
SOCORRO Approach: Data, AI and demonstration
The current SOCORRO project focusses on uniform corrosion of unprotected steel in contact with water. Once validated, the approach can also be applied to coated assets, cathodic protection, the occurrence of Microbiologically Influence Corrosion (MIC), etc. To gauge the corrosion risk presented by the environment, water quality parameters such as pH, Dissolved Oxygen, Temperature and Conductivity are measured. AI models are being developed to translate these measured parameters into corrosion risk over a period of time.
We are currently conducting laboratory experiments to collect training data for the AI models. The goal of the lab experiments is to simulate all possible environments to which real structures could be exposed. Both the water parameters and the corrosion rate are measured continuously using a carefully selected set of sensors. Lab experiments for both seawater and wastewater environments have been setup.
In addition to the lab experiments, 11 field demonstrations will be established for different types of applications, ranging over offshore foundations, ships and harbours to cooling water systems and wastewater treatment facilities. The data from these demonstrations will be used to finetune the measurements and models to adapt them to the different applications.
The collected and analysed data finally has to be presented to decision makers in a meaningful way. A visual interface to display the information and warnings is being developed based on calculated parameters like the instantaneous corrosion risk, accumulated corrosion risk and time averaged corrosion risk. In some instances, the system may also enable the decision maker to generate automated work orders based on warnings and alarms coming from these decision support tools.
Stay tuned for more detailed information on our progress and results!