Goal

The goal of this project is to develop a Proof of Concept model to determine the operational condition of the CAN RIG Top Drive 1250AC – 500T AC and demonstrate that sensor data may be used to predict field failures with enough lead time to take preventative action.

Model Brief

Predictive model that identifies potential failures for top drive system and/or components (Wash Pipe, Gears, Bearing Failure and Structural Failure). Model targets failure predictions with a minimum threshold of 8 hours.

Architecture and Technology

Built on Azure platform using IOT Hub to pull data from on premise sensors. The sensor data was cleansed and stored in Azure HDInsight (HDInsight 4.0). Azure Machine learning Model hosting in Azure Machine Learning Services, exposed via web services. The data visualization was built on PowerBI.