Multivariate Time Series Data usable for Time Series Segmentation and Time Series Classification. Each sample represents the multi-phased End-of-Line-Testing cycle of one hydraulic pump (evolution of 9 sensors). For confidentality reasons, the data were normalized (standard-score) and the sensor names anonymized.
The dataset was published in the context of one of our research articles:
Gaugel, S.; Reichert, M.: PrecTime: A Deep Learning Architecture for Precise Time Series Segmentation in Industrial Manufacturing Operations, 2023
Folder Structure:
Data/ Generation A/ Corresponding Generation. We have data from 2 generations (Generation A, Generation B) control type/ Coresponding pump control type. We have 3 different control types (Direct Control (DC), Proportional Control (PC), Speed-based Control (SC) ) version/ Coresponding version of specific control type. Only relevant for Generation A DC-pumps, where we have 3 version (V35, V36, V38) zip-file: zip file containing all samples of the specific version. One Sample has the following name format: "Pump_"+[Generation]+"_"+[Control Type]+[Version]+"_"+[ID]+".csv"
Generation B/ Corresponding Generation. We have data from 2 generations (Generation A, Generation B) control type/ Coresponding pump control type. We have 3 different control types (Direct Control (DC), Proportional Control (PC), Speed-based Control (SC) ) zip-file: zip file containing all samples of the specific version. One Sample has the following name format: "Pump_"+[Generation]+"_"+[Control Type]+[Version]+"_"+[ID]+".csv"
Each sample contains a time-series with 11 channels (data collected at 100 Hz frequency):
- Time-Index (in seconds)
- Values of Sensors 1-9 (standard-score normalized, no units for confidentality reasons)
- State Label (integer-encoded, meaning not further specified for confidentality reasons)
Publication 1: "PrecTime: A Deep Learning Architecture for Precise Time Series Segmentation in Industrial Manufacturing Operations" (2023)
Authors: Gaugel, S.; Reichert, M.
The subset referenced in the paper is found in the following path:
V35: Data/Generation A/DC/V35 V36: Data/Generation A/DC/V36 V38: Data/Generation A/DC/V38
Publication found at https://www.sciencedirect.com/science/article/pii/S0952197623002622
Publication 2: "Industrial Transfer Learning for Multivariate Time Series Segmentation: A study on the example of hydraulic pump testing cycles" (2023)
Authors: Gaugel, S.; Reichert, M.
The subsets referenced in the paper are found in the following paths:
DC-V35: Data/Generation A/DC/V35 DC-V36: Data/Generation A/DC/V36 DC-V38: Data/Generation A/DC/V38 SC: Data/Generation A/SC/V12 PC: Data/Generation A/PC/V23
Publication found at https://mdpi-res.com/d_attachment/sensors/sensors-23-03636/article_deploy/sensors-23-03636.pdf?version=1680246170
Publication 3: "Supervised Time Series Segmentation as Enabler of Multi-Phased Time Series Classification: A Study on Hydraulic End-of-Line Testing" (2023)
Authors: Gaugel, S.; Wu, B.; Anand, A.; Reichert, M.
The subset referenced in the paper is found in the following paths:
Generation B: Data/Generation B/DC/
Publication found at https://ieeexplore.ieee.org/document/10218185
Publication 4: "Data-Driven Multi-objective Optimization of Industrial End-of-Line Testing Cycles via Wrapper Feature Selection" (2024)
Authors: Gaugel, S.; Reichert, M.
The subset referenced in the paper is found in the following paths:
Generation B: Data/Generation B/DC/
Publication found at https://www.sciencedirect.com/science/article/abs/pii/S175558172400004X
The dataset created for the research located in the directory data are licensed under a Creative Commons Attribution 4.0 International License (CC-BY-4.0) .