Collaborative entropy perception application on compressor stall instability with topological visualization

To detect the initial instability in compressor, this study introduces a signal analysis application of system Kolmogorov entropy synergistically with multi-node integration. It employs the topological visualization with physical embedding to monitor the real-time status of the system. Based on multiple sensor signals, the proposed method extracts the intrinsic mode functions of system and maps into a high-dimensional space to derive the system high-dimensional Kolmogorov entropy. The inception of system instability can be identified by a significant increase in entropy. Simultaneously, the Mapper algorithm is utilized to visualize the topological structure of the instantaneous states of system in the high-dimensional space. Distinct topological structures correspond to different system states. Finally, the temporal-spatial stability of system is evaluated by combining the high-dimensional Kolmogorov entropy of system modal features with the topological structural diagrams. The results demonstrate that this method effectively partitions compressor stability into three distinct state spaces: stable, transition, and unstable. To evaluate the reliability of the synergistic sensing method, this paper introduces the improved Hilbert–Yellow transform, which accurately identifies the stall inception from the perspective of instantaneous frequency. By the collaborative perception strategy, the critical boundaries of transition and unstable states are marked by 285r and 5r in advanced, respectively. The modal feature measurement algorithm based on topological structures provides an approach to synergistic sensing for stability identification in nonlinear system.