Performance Optimization in Process Compressors
Challenge
Unplanned downtimes lead to reactive and time-based maintenance which can be costly and inefficient
Difficult to quantify metrics that correlate to valve health
Solution
Use first principles to predict the efficiency and power in real-time
Develop ML model to predict the anomalies in real-time leading to sub-optimal operations
Email notifications for real-time deviations
Recommendations to close the loop for optimal operations
Result
2-5% reduction in energy consumption
45-50% Reduction in unplanned downtimes