Sliding limited penetrable visibility graph for establishing complex network from time series.

Area of Science:

• Network science
• Time series analysis
• Railway engineering

Background:

• Time series data often contains noise and exhibits complex dynamics.
• Traditional network models may not effectively capture the temporal characteristics of time series.
• Diagnosing faults in railway track circuits, specifically compensation capacitors, requires robust analytical methods.

Purpose of the Study:

• To introduce a novel network modeling approach, sliding window limited penetrable visibility graph (SLPVG), for time series analysis.
• To apply the SLPVG model for the accurate diagnosis of compensation capacitor faults in jointless track circuits.
• To evaluate the effectiveness of network topological indicators, such as betweenness centrality, in fault detection.

Main Methods:

• Developed the sliding window limited penetrable visibility graph (SLPVG) algorithm to convert time series into network structures.
• Analyzed the degree distribution of various time series to validate SLPVG's ability to capture dynamic characteristics.
• Integrated fault characteristics of compensation capacitors with network topological indicators, focusing on betweenness centrality.
• Conducted experiments to test the SLPVG model's performance in identifying compensation capacitor faults.

Main Results:

• The SLPVG model effectively captures the dynamic characteristics of time series with low computational complexity.
• Betweenness centrality was identified as a key indicator for accurately reflecting the fault status of compensation capacitors.
• The proposed model achieved a high accuracy rate of 99.1% in identifying compensation capacitor faults in jointless track circuits.

Conclusions:

• The SLPVG model offers a simple, efficient, and accurate method for analyzing long time series dynamics.
• This approach provides a novel perspective for diagnosing compensation capacitor faults in jointless track circuits.
• The study demonstrates significant practical implications for advancing research in time series analysis and railway fault detection.