This research evaluates various methods to restore missing data in smart meter readings, ensuring reliable energy consumption analytics.

• Benchmark study comparing statistical, machine learning, and time series foundation models for data imputation
• Addresses nonlinear patterns in smart meter data that conventional techniques struggle with
• Enhances data integrity by resolving gaps caused by sensor failures and transmission errors
• Improves grid reliability by providing more accurate consumption analyses and predictions

For engineering teams, this research offers practical solutions to prevent technical and economic inefficiencies in smart grid operations while maintaining data quality as complexity and volume grow.

Bridging Smart Meter Gaps: A Benchmark of Statistical, Machine Learning and Time Series Foundation Models for Data Imputation