Models for predicting resistance trends

A system for predicting long-term resistance trends at cathodic protection sites using site-specific metadata for preventive maintenance planning, comprising:

• at least one computer;
• and,
• at least one non-transitory computer-readable medium storing computer-readable instructions that, when executed by the at least one computer, cause the at least one computer:
• obtain, a plurality of historical resistance trends from a plurality of installed rectifier sites;
• obtain a plurality of site-specific metadata features for the plurality of installed rectifier sites, wherein the site-specific metadata features are physical, environmental, and installation characteristics of each installed rectifier site of the plurality of installed rectifier sites and include at least soil characteristics, pipeline properties, and environmental data, the site-specific metadata features being absent voltage, current, and resistance measurements at any rectifier site of the plurality of installed rectifier sites;
• label each historical resistance trend of the plurality of historical resistance trends as one of a plurality of historic resistance trend classifications;
• input, into a machine learning algorithm, the labelled historical resistance trends and the site-specific metadata features of the plurality of installed rectifier sites to train a classification model;
• generate, by the classification model, a predicted resistance trend classification for a rectifier site in response to the site-specific metadata features for a rectifier site being input into the classification model, and independent of voltage, current, and resistance measurements at any rectifier site of the plurality of installed rectifier sites;
• and
• schedule predictive maintenance for the rectifier site based on the predicted resistance trend classification.

At least one non-transitory computer-readable media storing computer-readable instructions that, when executed by at least one computer, cause the at least one computer to predict rectifier resistance trends comprising:

• obtaining, by at least one computer, a plurality of historical resistance trends from a plurality of installed rectifier sites;
• obtaining, by the at least one computer, site-specific metadata features for the plurality of installed rectifier sites, wherein the site-specific metadata features are physical, environmental, and installation characteristics of each installed rectifier site of the plurality of installed rectifier sites and include at least soil characteristics, pipeline properties, and environmental data, the site-specific metadata features being absent voltage, current, and resistance measurements at any rectifier site of the plurality of installed rectifier sites;
• labelling each historical resistance trend of the plurality of historical resistance trends as one of a plurality of historic resistance trend classifications;
• inputting into a machine learning algorithm the historical resistance trends and the site-specific metadata features of the plurality of installed rectifier sites to train a classification model;
• generating, by the classification model, a predicted resistance trend classification for a rectifier site in response to site-specific metadata for the rectifier site being input into the classification model, and independent of voltage, current, and resistance measurements at any rectifier site of the plurality of installed rectifier sites;
• and
• scheduling preventive maintenance for the rectifier site using the predicted resistance trend classification.

A computer-implemented method of predicting rectifier site resistance trends comprising the steps of:

• receiving, by at least one computer, rectifier site metadata features specific to a rectifier site, wherein the rectifier site metadata features are physical, environmental, and installation characteristics of each installed rectifier site of the plurality of installed rectifier sites and include at least soil characteristics, pipeline properties, and environmental data, the rectifier site metadata features being absent voltage, current, and resistance measurements at any rectifier site of the plurality of installed rectifier sites;
• and,
• generating, by the at least one computer, a predicted resistance trend classification for the rectifier site by inputting the rectifier site metadata into a classification model trained to:
• receive, as input, the rectifier site metadata, independent of voltage, current, and resistance measurements at the rectifier site, and
• output a predicted resistance trend classification for the rectifier site based on the received rectifier site metadata;
• and
• scheduling preventive maintenance for the rectifier site using the predicted resistance trend classification.