Corrosion Predictions

Key Challenges for corrosion prediction

  • Non-linearity and inertia of industrial processes,​
  • Unobservable state variables,​
  • Low number of observations available due to low frequency of laboratory tests,​
  • Measurement errors and inaccuracies in laboratory results,​
  • Irregularities resulting from structural changes (changes in dependencies between variables), ​i.e. caused by charge fluctuations, installation loads and other factors

Benefits from the corrosion prediction

  • Forecasts of degradation mechanisms​
  • Increased safety and availability of installations​
  • Optimization and automation of maintenance​
  • Systematic, continuous growth of know-how and improvement of corrosion management solutions

Client’s Challenges

The corrosion rate for critical assets / lines is measured and the time of failure is predicted but no correlation with corrosion-promoting agents is discovered nor controlled actions with predictable results are taken

Proposed Solution

IoT sensors data sourcing with Advanced analytics and Artificial Intelligence models are used to discover the correlations between the quality of input and installations’ corrosion. These models are employed within the CMS to leverage slight modifications of input, addition of inhibitors or adjustments in process parameters (flow, pressure, temperature, etc.) to significantly reduce corrosion rates without unwanted impact on final product.

Results

Overall Equipment Effectiveness is significantly improved without modifications to the installation by reducing outages and need for spares​

Small, targeted changes result in large increase of corroding elements lifespan