Multiplex multivariate recurrence network from multi-channel signals for revealing oil-water spatial flow behavior

Area of Science:

• Fluid Dynamics
• Complex Systems Analysis
• Sensor Technology

Background:

• Understanding spatial dynamical flow behaviors in oil-water mixtures is crucial due to their complexity and industrial significance.
• Existing methods face challenges in comprehensively analyzing multi-channel signals from layered sensor systems.

Purpose of the Study:

• To develop a novel method for analyzing spatial dynamical flow behaviors in oil-water systems.
• To effectively fuse and interpret multi-channel signals from a double-layer distributed-sector conductance sensor.

Main Methods:

• Designed a double-layer distributed-sector conductance sensor for capturing spatial flow information.
• Developed a multiplex multivariate recurrence network (MMRN) based on recurrence network theory to fuse multi-channel signals.
• Derived projection networks from MMRNs and utilized average clustering coefficient and spectral radius for quantitative characterization.

Main Results:

• The proposed network measures (average clustering coefficient and spectral radius) are highly sensitive to changes in oil-water flow states.
• The distribution of these network measures effectively reveals the underlying spatial dynamical flow behaviors of different flow patterns.
• Demonstrated the capability of the method to analyze multi-channel signals from multi-layer sensor systems.

Conclusions:

• The novel MMRN method provides an efficient approach for analyzing complex spatial dynamics in oil-water flows.
• Network measures derived from MMRNs offer sensitive indicators for characterizing distinct flow patterns.
• This work advances the analysis of multi-channel data from advanced sensor systems in fluid dynamics.