# Milling Tool Wear Time Series across Three Machine Tools (Flank Wear VB, Forces, Spindle Torque)

## Facts

- Provider: Stiehl et al. (original); Parquet conversion by alpha-by-beta
- License: CC-BY-4.0 (https://data.mendeley.com/datasets/zpxs87bjt8/3)
- Size: 3 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/alpha-by-beta/Multivariate_time_series_data_of_milling_processes_with_varying_tool_wear_and_machine_tools
- Access: public, readable without an account
- DOI: https://doi.org/10.17632/zpxs87bjt8.3
- Category: Machine Learning & AI
- Source: https://doi.org/10.1016/j.dib.2023.109574
- Page: https://scigantic.com/archive/3d0491ec-166c-5b10-a109-f8a4812a8c87
- Keywords: Machine Learning & AI, time series, Parquet, Hugging Face

## About

Labeled multivariate time series of shoulder milling with progressive tool wear on three different 5-axis milling centers, for tool wear estimation, tool life prediction and cross-machine transfer learning. Nine solid carbide end mills (4 edges, TiN-TiAlN coated, same model, different batches) were worn from new to end of life (flank wear land VB about 150 micrometers) in 3-axis shoulder milling of cast iron 600-3/S, with VB measured by digital microscope at 100x. Process forces were recorded with a dynamometer at 25 kHz; spindle and feed drive torques or forces, tool position and position control deviation came from the machine controls at 500 Hz. Labels per run are machine, tool, run number, cumulated tool contact time and wear. The original 6,418 files are flattened here into a single dataset.parquet (about 23 GB) readable with pandas, polars or pyarrow. Original data by Stiehl et al. on Mendeley Data under CC BY 4.0; cite doi:10.17632/zpxs87bjt8.3 and the Data in Brief article doi:10.1016/j.dib.2023.109574.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("alpha-by-beta/Multivariate_time_series_data_of_milling_processes_with_varying_tool_wear_and_machine_tools")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `dataset.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/3d0491ec-166c-5b10-a109-f8a4812a8c87
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/3d0491ec-166c-5b10-a109-f8a4812a8c87
- Over MCP: `get_schema_card` with id `3d0491ec-166c-5b10-a109-f8a4812a8c87` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Time series datasets](https://scigantic.com/topics/time-series)
- [Machine Learning & AI datasets](https://scigantic.com/topics/machine-learning-and-ai)
