Abstract
The rectifier groundbed is a key component of an Impressed Current Cathodic Protection (ICCP) system in that it is consumed over time. The rate of consumption is based on many contributing factors including anode mass and material, surrounding soil properties, seasonal effects and rectifier operating conditions. Accurately forecasting the remaining service life of an anode groundbed has proven challenging for pipeline operators in the past, faced with the decision to replace the groundbed well before the anodes have been consumed or waiting until groundbed failure and risking a time period where the pipeline is no longer cathodically protected.
We present a statistical approach to this problem, using remote-monitoring data and advanced data analytics techniques. A machine learning model is developed to classify rectifier readings as in their nominal phase or approaching the end of service life. We present and discuss what factors have the most influence on the model and discuss wider applicability on a large scale of rectifier datasets.
