The digital production twin provided additional no-measured values utilized in compressor performance optimization incl. energy consumption and troubleshooting
BPT deployed a high-fidelity digital production twin for a centrifugal compression facility.
The twin solution used all available real-time sensors as input, including automatically compositional update to each source and their rates. The sensor values were quality-checked holistically and corrected by the reconciliation solver calculating the best-fit values. Faulty and bad sensor values were replaced.
The twin also provided a complete dataset of augmented soft sensing values, powered by the 1st principle rigorous thermodynamic simulator model and high-fidelity compressor stage application with corrected performance map (impeller-by-impeller).
The digital compressor twin, driven by corrected sensor values and the additional soft sensing data, such as vapour compositions and properties, facilitated compressor performance optimization including energy consumption and operational troubleshooting.
Further fidelity in the compressor performance optimization were achieved when integrating machine learning application with the physical based BPT Twins.