Electricity Transformer Oil Temperature and Load Data from China

Featuring hourly and 15-minute frequency measurements of electricity transformer oil temperature and six power load features from two counties in China, collected over two years. The data is split into 12 months for training, 4 for validation, and 4 for testing.

Use Cases

• Forecast oil temperature using the six power load features as inputs for multivariate time series models.
• Analyze the relationship between specific power load features and oil temperature to identify key operational drivers.
• Train univariate time series models on the oil temperature sequence for predictive maintenance applications.
• Validate forecasting models using the predefined 12/4/4 month train/val/test split.

Strengths

• Data collected at two high frequencies (hourly and 15-minute) over a two-year period.
• Includes a predefined 12-month training, 4-month validation, and 4-month testing split.
• Contains seven variables per data point: one target (oil temperature) and six explanatory power load features.

Limitations

• The exact number of rows and sample size are unknown, making it difficult to assess scale.
• Geographic coverage is limited to only two counties in China, which may limit generalizability.
• Data freshness is unclear beyond the last platform update date.

Provenance

Source

ETDataset

Collection Method

Collected from electricity transformers in two counties.

Time Range

Two years of data (specific years not provided).

Freshness

Last updated on the platform on 2024-01-18.

Geography

Two counties in China.