The day is going well, and the plant is running normally. Then suddenly, you have a gas leak. Your first question is, where is the leak?

Then other questions follow.

1.     How to stop this leakage?

2.    Why did this leakage happen?

3.    How to safely shutdown the plant to avoid an accident?

After controlling the situation, a team will be formed. Their job will be to identify the reason for the leak, amount of loss the plant incurred, and in the worst-case situation, count the number of deaths.

Corrosion – it is a slow poison to plant equipment and pipelines. Its effects are often only apparent when the system has already sufficiently corroded, and destruction is either about to take place or already occurred.

Yokogawa Electric Corporation (here after will be referred as Yokogawa) and KBC (a Yokogawa company) developed a solution to this problem. Predict corrosion and inform plant people in advance so that operators can take corrective actions to avoid the disaster.

Our answer to this problem is a hybrid solution that combines Artificial Intelligence (AI) based on Yokogawa Machine Learning techniques and KBC’s Petro-SIM® simulation model with the OLI Alliance Engine*.   

We provide an accurate corrosion condition prediction system by combining Artificial Intelligence/Machine Learning (AI/ML) technology and Petro-SIM first principles-based simulation technology and the OLI electrolyte simulation package. This allows operators to monitor and manage systems, such as Health Index and Corrosion rate. 

Based on the close integration with rigorous simulation technology and AI/ML technology, this system can also predict future corrosive and case. Operators can then become proactive with corrosion management. The information this system provides allows operators to develop mitigation plans to manage future corrosion with predictive maintenance.

Chemistry Involved

Whenever chloride salts are mixed with water, they form hydrochloric acid (HCl). Amines are added to neutralize the pH of the process streams, however amine chlorides at certain temperatures form an undesired solid which can lead to under deposit corrosion.

The OLI Alliance Engine integration with Petro-SIM

Electrolyte ionic modeling is necessary for accuracy in studies which require rigorous prediction of chemical reactions and phase behavior of complex chemical mixtures in water under varying temperatures, pressures, and ionic strengths. (Note: OLI requires separate license)

Application

Ionic modeling – that is, electrolyte thermodynamics models can determine the conditions when the amine – HCl reaction forming an undesired solid takes place and thus avoid operating at that temperature. Amine chloride solids can lead to under-deposit corrosion.

Petro-SIM with the OLI engine determines those factors which are triggering the corrosion in the system.

Solution

Yokogawa’s AI team developed the AI/ML model which can be configured for corrosion prediction by learning processes from several historical operations data.

The application of AI/ML model

1.     The AI/ML model is placed on-line, to take the current operating data at real time from plant through DCS.

2.     The AI/ML model can predict the future problem by showing the health Index (Positive is OK indication, Negative is Bad indication for future).

3.     The AI/ML model also provides top three “cause variables” which are directly related to the cause of corrosion.

Finally, the On-line Digital Twin model (Petro-SIM with OLI Engine) is used to simulate the top three “cause variables” (found in step 3 above) to find out the mitigation plan to avoid future corrosion problems.

This Digital Twin model is a combination of Petro-SIM with OLI simulation and the AI/ML model; With the help of this Digital Twin Model used for monitoring and diagnostics of the asset performance, a plant operator can make the right decision at right time to mitigate future corrosion problems and help the plant to avoid mishaps.

*The OLI Alliance Engine is a product of OLI Systems Inc