Using Predictive Analytic Software Solution for Idendifying Equipment Failure in Advance; the University of Texas at Austin

The University of Texas at Austin is well known for its quality of education and research. In order to maintain a high standard of education, the uninterrupted and cost-effective power, steam and cooling supply plays an important role. To meet this need, the University of Texas at Austin – Utilities and Energy Management Department decided to adopt an innovative predictive maintenance based solution. A technology wise Predictive Maintenance approach promises cost savings and improvement in both stability and performance compared to a routine or time based Preventive Maintenance program. UT-UEM has deployed the Predictive Analytic Software Solution; PHI (Plant Health Index) to improve Operational Excellence by using the early detection feature of the system that captures anomalies in the behaviours of sensors, equipment and the operation of process plant systems. PHI also helps analyse process uncertainty using statistical learning by modelling the operational history in a simple setup. This paper explains how UT-UEM achieves its goals and identifies hidden problems in advance before they become catastrophic failures with the help of PHI. The case studies shows how UT-UEM used the Predictive Analytic Software Solution to provide an early warning of anomalies that could result in system and operational failures in its systems. The studies also shows how UT-UEM identified the abnormal behaviour and hidden problems from the sensor level to the associated equipment level for Gas Turbines, Steam Turbines and Boilers systems.