Condition Monitoring for Packaging Industry dataset (CoMoPI)

<p><strong>Condition Monitoring for Packaging Industry dataset (CoMoPI)</strong></p> <p>CoMoPI dataset contains data from eight industrial packaging machines:</p> <table> <thead> <tr> <th> <p><strong>serial</strong></p> </th> <th> <p><strong>Starting Date</strong></p> </th> <th> <p><strong>Ending Date</strong></p> </th> </tr> </thead> <tbody> <tr> <td> <p>A001</p> </td> <td> <p>2022-08-10 03:45:00+00:00</p> </td> <td> <p>2023-01-09 07:00:00+00:00</p> </td> </tr> <tr> <td> <p>A005</p> </td> <td> <p>2022-09-19 13:15:00+00:00</p> </td> <td> <p>2023-01-09 15:45:00+00:00</p> </td> </tr> <tr> <td> <p>B002</p> </td> <td> <p>2022-07-14 14:00:00+00:00</p> </td> <td> <p>2023-01-09 15:50:00+00:00</p> </td> </tr> <tr> <td> <p>B005</p> </td> <td> <p>2022-10-04 12:55:00+00:00</p> </td> <td> <p>2023-01-05 01:55:00+00:00</p> </td> </tr> <tr> <td> <p>C003</p> </td> <td> <p>2022-12-15 06:50:00+00:00</p> </td> <td> <p>2023-01-03 16:35:00+00:00</p> </td> </tr> <tr> <td> <p>C004</p> </td> <td> <p>2022-07-18 07:55:00+00:00</p> </td> <td> <p>2022-12-28 08:10:00+00:00</p> </td> </tr> <tr> <td> <p>E002</p> </td> <td> <p>2022-07-14 14:20:00+00:00</p> </td> <td> <p>2022-12-03 14:00:00+00:00</p> </td> </tr> <tr> <td> <p>E004</p> </td> <td> <p>2022-11-23 15:25:00+00:00</p> </td> <td> <p>2023-01-09 10:00:00+00:00</p> </td> </tr> </tbody> </table> <p>The dataset provides sensor measurements related to a specific module involved in the watertight closure of packages. It also provides all alarms and warnings generated by the packaging equipment.</p> <p><strong>Dataset description</strong></p> <p>For the sake of simplicity, the dataset is divided into three files, one for sensor measurements, one for alarms and one for warnings.</p> <p><em>Sensor measurements</em></p> <p>This data is included in the file called&nbsp;<em>industrial_dataset_sensors_10m_agg.csv</em>. Each row corresponds to average sensor value in a 10-minute window. Beyond the machine identifier (&#39;_serial&#39;) and timestamp (&#39;_time&#39;), the following measurements are available: &#39;AE&#39;, &#39;BE&#39;, &#39;AF&#39;, &#39;BF&#39;, &#39;APP&#39;, &#39;BPP&#39;, &#39;AP&#39;, &#39;BP&#39;, &#39;ALE&#39;, &#39;BLE&#39;, &#39;ALP&#39;, &#39;BLP&#39;, &#39;ADS&#39;, &#39;BDS&#39;, &#39;AES&#39;, &#39;BES&#39;. Suffixes A_ and B_ relate to the two elements that can perform a specific operation required by the watertight closure of packages. For confidentiality reasons, it is not possible to provide further details.</p> <p><em>Alarm measurements</em></p> <p>This data is included in the file called&nbsp;<em>industrial_dataset_alarm_10m_agg.csv</em>, according to the following schema:</p> <ul> <li>&#39;_serial&#39; : &#39;machine unique identifier&#39;</li> <li>&#39;_time&#39; : &#39;timestamp&#39;,</li> <li>&#39;AL_1&#39; : &#39;Counter of AL_1 alarms&#39; ,</li> <li>...</li> <li>&#39;AL_123&#39; : &#39;Counter of AL_123 alarms&#39;</li> </ul> <p><em>Warning measurements</em></p> <p>This data is included in the file called&nbsp;<em>industrial_dataset_warnings_10m_agg</em>.csv_, according to the following schema:</p> <ul> <li>&#39;_serial&#39; : &#39;machine unique identifier&#39;</li> <li>&#39;_time&#39; : &#39;timestamp&#39;,</li> <li>&#39;WR_1&#39; : &#39;Counter of AL_1 alarms&#39; ,</li> <li>...</li> <li>&#39;WR_358&#39; : &#39;Counter of WR_358 alarms&#39;</li> </ul> <p><strong>Condition monitoring</strong></p> <p>Alarms &#39;AL_53&#39; are &#39;AL_54&#39; are related to faulty conditions of the components and can be used as prediction target. The following alarms are generated by the target module: &#39;AL_17&#39;, &#39;AL_18&#39;, &#39;AL_40&#39;, &#39;AL_41&#39;, &#39;AL_42&#39;, &#39;AL_43&#39;, &#39;AL_45&#39;, &#39;AL_46&#39;, &#39;AL_47&#39;, &#39;AL_48&#39;, &#39;AL_49&#39;, &#39;AL_50&#39;, &#39;AL_51&#39;, &#39;AL_52&#39;, &#39;AL_53&#39;, &#39;AL_54&#39;.</p> <p><strong>Anonymization procedure</strong></p> <p>The anonymization procedure is the following:</p> <ul> <li>To make it impossible to identify the specific piece of equipment, its serial number was replaced with a mock equipment_ID. Moreover, no additional information is provided about the machine, its location and the processed products.</li> <li>Sensor measurements were renamed and rescaled to remove all information related to actual working setup.</li> <li>Alarms and warnings underwent an anonymization process: no description is provided, and the original alarm codes were mapped to mock codes.</li> </ul>