IIoT Metalworking Fluid Degradation — Real-World Physicochemical Sensor Monitoring [Multi-Month]
Abstract
"Real-world IIoT multivariate time series dataset tracking physicochemical degradation of metalworking fluid over several months. Includes imputed benchmark variants for 5 methods. CSV format. Designed for predictive maintenance and anomaly detection research in manufacturing."
Description
Overview
This dataset presents a real-world IIoT multivariate time series recording the physicochemical degradation of an industrial metalworking fluid (MWF) collected continuously over several months from a test tank at an industrial facility in Spain. Metalworking fluids are critical to CNC machining operations and their degradation directly impacts tool life, surface finish, and worker health.
The dataset was created within the SmarTaladrine project and is notable for its real operational provenance — not simulated — along with the inclusion of deliberately injected missing data patterns and five benchmark imputation outputs (KNN, HybridKCL, LSTM-VAE, and pre-trained and fine-tuned MOMENT foundation model). This makes it a dual-purpose resource for both predictive maintenance model development and time-series imputation benchmarking.
Published in 2025 through ScienceDirect Data in Brief, the dataset is accompanied by a full methodology article providing context for the physicochemical variables monitored, the operational conditions, and the imputation protocols applied.
Column Schema
| Column | Description |
|---|---|
| timestamp | Continuous monitoring timestamp over multi-month deployment. |
| pH | pH of the metalworking fluid (critical degradation indicator). |
| conductivity | Electrical conductivity measurement. |
| concentration | Fluid concentration level. |
| temperature | Fluid temperature during operation. |
| turbidity | Fluid turbidity measurement (contamination indicator). |
| imputed_KNN | KNN-imputed values for missing data segments. |
| imputed_LSTM_VAE | LSTM-VAE imputed values for missing data segments. |
| imputed_MOMENT | Foundation model imputed values (pre-trained and fine-tuned). |
Key Statistics
- Data Type: Real-world IIoT continuous monitoring (not simulated)
- Duration: Several months of uninterrupted operation
- Imputation Benchmarks: 5 methods included per variable
- File Format: CSV
- Published: 2025 (ScienceDirect Data in Brief)
- Source Facility: Industrial site in Spain (SmarTaladrine project)
Use Cases
- Predictive maintenance modelling for industrial fluid systems and CNC machining
- Anomaly detection in physicochemical sensor streams
- Benchmarking time-series imputation algorithms on real IIoT data with missing values
- Sustainable manufacturing analytics and fluid lifecycle management
Source & Attribution
The dataset was published as a data article in Data in Brief (Elsevier / ScienceDirect) in 2025, developed by researchers from the University of Burgos, Spain, within the SmarTaladrine project. It is publicly accessible through the ScienceDirect article page.
View Data Structure
To explore column names, data types, and sample rows, visit the official dataset page on Research Paper.
Preview on Research PaperCite This Dataset
Carlos Cambra, Félix Movilla, & Félix {de Miguel (2025). IIoT Metalworking Fluid Degradation — Real-World Physicochemical Sensor Monitoring [Multi-Month]. Data in Brief. [Dataset]. Research Paper. https://doi.org/10.1016/j.dib.2025.112020
Source: Research Paper (2025) · DOI: https://doi.org/10.1016/j.dib.2025.112020
Indexed by IoTDataset.com on Apr 10, 2026
Ready to Start Your Research?
Download this dataset directly from the official repository and start building your next breakthrough project.