During the deployment and retrieving of the sensor nodes, two notable incidents happened that affected the occupants negatively. Two more incidents affected only the data acquisition In the first incident, a freezer stopped working in the night after the sensor deployment. For the installation of a router node, I unplugged the freezer, inserted a multi-plug, plugged in the freezer and the router node. After the installation of all sensors, I notified the building owner of the procedure and even opened the freezer door to demonstrate that it was still working. To me, it is unclear if the unplugging was the cause of the freezer thawing or if it was just an unlucky coincidence. However, the building owner was able to eat all the contents in a timely fashion. So besides the inconvenience, there was no damage. The freezer continued to work after the incident.

The second incident happened during the removal of a sensor network. The gateway of the WSN was plugged into a multi-plug together with the building owner’s modem. To retrieve the multiplug, I unplugged the modem, removed the multiplug and plugged the modem in again. However, the modem refused to work again. The building owner was working from home and was less than amused. The modem was replaced by the internet provider without charge. The internet provider stated that unplugging and re-plugging the modem is acceptable and should not corrupt the modem. Nevertheless, it caused some inconvenience for the case study participants, which was compensated with a bottle of wine and some chocolate.

In one instance, the heating system reached its end of life and was replaced by a new one. In this process, the sensor nodes attached to the heating system were discarded together with the heating system. This had also the consequence that for this building, we were only able to acquire a limited amount of heating data. All other sensors nodes could be retrieved intact.

After the measurement campaign, I left the sensor nodes operating as they were in our office space, in order to assess the battery lifetime. During this time, a fourth incident happened. The admin from our institution took the webserver down because the responsible person from our time left the institution. This took all sensors offline and would have been very inconvenient during the campaign. However, after a few emails, the server was up again, and all sensor nodes would resume reporting measurement data.

In the beginning, we had the concern that pets might interfere with the cables from the sensor nodes. However, pets never caused an issue with the sensor network.