Information-theoretic grid topology reconstruction using low-precision smart meter data.

Area of Science:

• Electrical Engineering
• Network Science
• Data Science

Background:

• Power grid topology is crucial for state estimation and stability.
• Existing data-driven methods lack defined measurement quality requirements.
• Understanding data fidelity is key for robust topology reconstruction.

Purpose of the Study:

• To determine the minimum data fidelity needed for distribution grid topology reconstruction using voltage measurements.
• To analyze the sensitivity of topology reconstruction to measurement quantization, precision, and sampling frequency.

Main Methods:

• Utilized an information-theoretic approach with the Chow-Liu algorithm to generate maximum spanning trees based on mutual information.
• Conducted a sensitivity analysis on data bit-depth, significant digit truncation, time-window length, and mutual information estimators.
• Validated the approach using IEEE test cases (MATPOWER) and GridLAB-D time-series data.

Main Results:

• Successful grid topology recovery was achieved with 8-bit quantized data and millivolt-level precision.
• Reconstruction accuracy significantly degraded with downsampling intervals over 20 minutes or limited data duration.
• Established an optimistic theoretical lower bound for data quality in structural inference.

Conclusions:

• High-precision instrumentation may not be essential for power grid structural inference under ideal conditions.
• The findings provide a baseline for evaluating real-world noisy smart meter data.
• This research supports the development of hybrid approaches incorporating engineering priors for topology reconstruction.