The Interconnected Tank System Dataset (InTaSet) is a comprehensive benchmark dataset for nonlinear system identification and sensor-fault detection, based on real-time measurements from a physical interconnected quadruple-tank testbed operating in both open-loop and closed-loop configurations.
The main folder InTaSet.zip contains two subfolders: (1) dataset for system identification (InTaSet-ID) and (2) dataset for sensor-fault detection (InTaSet-FD).
Within the InTaSet-ID sub-folder, four testbed configurations are provided. Each configuration uses different training and testing datasets.
Within the InTaSet-FD sub-folder, it contains 4-subsubfolders: (1) clean (closed-loop operating modes), (2) sensor-fault to Tank 1, (3) sensor-fault to Tank 3, and (4) sensor-fault to Tank 1 and 3.
If you use the above dataset for evaluating system identification and sensor-fault detection frameworks, please cite our EWRI conference article below:
Putri, S. A., Daniela, V., Raza, N., Iwakin, O., & Moazeni, F. (2025). InTaSet: A Benchmark Dataset for Data-Driven System Identification and Fault Detection in an Interconnected Water System. World Environmental & Water Resources Congress 2026 (EWRI 2026), Mobile, Alabama. [DOI will be provided soon]
Or you can directly cite our dataset as follows:
Putri, S. A., Villacrés, D., Raza, N., Iwakin, O., & Moazeni, F. (2025). InTaSet: A Benchmark Dataset for Data-Driven System Identification and Fault Detection in an Interconnected Water System [Data set]. Zenodo. https://doi.org/10.5281/zenodo.17652851