We explore how rectifier voltage and current measurements can inform pipeline engineers and technicians on the health, performance and operation of their cathodic protection (CP) assets, and predict the future operation of existing and newly installed cathodic protection systems. We leverage years of data from monitoring units installed on CP rectifiers combined with site specific details describing the site and its CP system provided by pipeline operators to train a machine learning model.

The study includes current and historical data from hundreds of unique rectifier locations across Canada which have been historically monitored using a remote monitoring unit (RMU). RMU readings are analyzed and grouped by long term resistance trends. Contextual data is collected for each site. This data describes the cathodic protection relevant details of the site, including details of the pipe, rectifier, groundbed and soil.

A machine learning model has been developed which accepts the contextual data associated with the rectifier and will predict the long-term rectifier resistance trend.


Impressed current rectifiers are the backbone of a pipeline operator's cathodic protection (CP) systems. A rectifier's ability to protect a large length of electrically continuous pipeline considerably improves efficiencies and reduces material costs as compared to galvanic systems1, 2. However, like galvanic anodes, impressed current anodes are a consumable asset, and require replacement at the end of their service life to ensure that the rectifier can continue to adequately protect the pipeline.

The CP industry has relied on system resistance, commonly calculated using voltage and amperage data collected from rectifiers, as a leading indicator of rectifier groundbed performance3. However, much of this analysis is performed manually, irregularly, and on a limited data set of few assets, which has precluded the depth of insight that this method has been able to generate for practitioners4.

In the recent decades, advances in cellular and satellite communications have enabled remote monitoring of the electrical parameters associated with cathodic protection. This has allowed operators to gather data more frequently and save time by avoiding unnecessary visits to hard-to-reach sites. The next phase in technological improvement of CP system monitoring is analyzing large amounts of remotely gathered data and using machine learning algorithms to make predictions and gain insight from various related datasets4–6.

This content is only available via PDF.
You can access this article if you purchase or spend a download.