Real-time estimation of a multi-stage centrifugal compressor performance map considering real-gas processes and flexible operation

Abstract

This paper contributes to modeling and supervision of multi-stage centrifugal compressors coping with real-gas processes and steady to highly transient operating conditions. A novel dynamic model is derived, and the incorporation of the generic Lee-Kesler-Plöcker real-gas equation of state and its derivatives is presented. The model allows for embedding arbitrarily shaped performance maps, based on state-of-the-art polytropic change-of-state compressor characteristics. As the validity of these maps is a key issue for simulation and model-based monitoring, performance maps are treated as time-variant, and their shape is to be identified and monitored during operation. The proposed real-time map estimation scheme comprises an Unscented Kalman Filter and a newly proposed algorithm, referred to as Recursive Map Estimation. The combination yields a novel parameter and state estimator, which is expected to be superior if some parameters are characterized by a distinct operating point dependency. Two additional time-variant parameters are provided for monitoring: The first indicates the level of confidence in the local estimate, and the second points to drastic performance map alterations, which may be further exploited in fault detection. A modified reference simulation of a two-stage supercritical carbon dioxide compressor with known state trajectories, performance maps, and alterations demonstrates the successful application of the entire monitoring scheme, and serves for a discussion of the results.

Guide Vane Opening Prediction for Constant Speed Axial Blowers in Blast Furnace Ironmaking with Variation Information

“…The conventional simulation methods, such as the Greitzer model or compressor maps, use equations of fluid flow dynamics to describe system performance. [15][16][17][18] However, the ML-based models have been proved to be more accurate and faster than the computational fluid dynamics-based simulations. 12) Moreover, the air blower operated in practice makes it much more difficult to determine the suitable quantitative expressions as, for a generic value of GVO, there could be various combinations of air flow rate and pressure.…”