PreDist Dataset - Operational data of district heating substations labelled with faults and maintenance information

This dataset consists of operational data and labels based on incident reports and maintenance data of district heating substations of enercity Netz GmbH. The labels are available as a list of ‘disturbances’ as well as a list of fault reports including a short description and problem category, which can be used to develop (early) fault detection models for district heating substations. In addition, fault labels and monitoring potential were added to the reports where possible. The dataset is published together with the paper "Enabling Predictive Maintenance in District Heating Substations: A Labelled Dataset and Fault Detection Evaluation Framework based on Service Data", which explains the dataset in detail. When referring to this dataset, please cite the paper mentioned in the related work section. 

The PreDist dataset contains time series of 93 district heating substations from two manufacturers, M1 and M2, each time series spanning different lengths of time, depending on when the substation was ‘digitised’. Both sub-datasets contain a list of faults (based on incident reports), a list of disturbances (incident reports, and corrective and preventive maintenance tasks and activities), feature descriptions and a list of pre-defined ‘normal events’, which can be used in addition to the faults to evaluate normal behaviour models.

Changes in version 2:
A list of all substations with associated configuration type has been added to each manufacturer (file 'configuration_types.csv'). The readme file has been updated accordingly and includes a short description of the different configuration types present in the dataset.