Development of Predictive Maintenance Interface Using Multiple Linear Regression

The common practice in most of the oil and gas organizations is to use periodic and preventive (time-based) maintenance techniques. These time-based maintenance techniques apply checking on equipment at regular intervals to avoid failures. However, they can lead to high maintenance cost with either over maintenance or unscheduled downtime. This issue can be encountered by using predictive maintenance which predicts future equipment failures ahead of time. This method performs maintenance based on actual operating condition of equipment which reduces maintenance cost and eliminates the need for periodic maintenance. The purpose of this work is to develop a user-friendly Graphical User Interface (GUI) application based predictive maintenance data analytic interface with the implementation of Multiple Linear Regression predictive maintenance technique. Multiple data sets of parameters of Booster Compressor (BC) are used on the proposed GUI to determine the accuracy of future prediction through implementation of Multiple Linear Regression technique.