EAPBSdata: Energy consumption and bee dataset of five connected beehives of one high season in France

<strong>1- Introduction</strong> This dataset was developed in the context of our work [1] that focuses on the orchestration of tasks between edge devices and cloud servers, and the energy optimization of orchestration scenarios applied to smart beehives. [1] introduces an energy-aware precision beekeeping system capable of collecting bee-related data and its own energy consumption. This system was deployed during the 2022 beekeeping high season in 5 beehives, 3 of which located in Lyon, France, and the 2 others in Cachan, France. <strong>2- Description</strong> <strong>Hardware</strong> The energy-aware precision beekeeping system relies on the Raspberry Pis: One Raspberry Pi 3b+ that periodically wakes up, collects bee-related data, transfers them and shuts down. One Raspberry Pi Zero WH that constantly collects the energy consumption of itself, the Raspberry Pi 3b+ and the photovoltaic intake. The energy node of our system relies on a solar panel, current converter, and battery. The solar panel is a monocrystalline 30 W panel. The battery is a 20 000 mAh power bank. Between the two, the current is converted to 5 volts (the adequate voltage for the battery) with a DC/DC Step-Down 5 volts/3 amperes converter. The current sensors connected to the Raspberry Pi Zero are Grove ±5A DC/AC current sensors based on the Allegro ACS70331 current sensor, which is based on giant magneto-resistivity sensing technology. This sensor is selected because of its adapted sensing range (0 to 5 A) and its wide temperature operating range (-40°C to 85°C), as the later-introduced setup operates in a sealed case in an outdoor environment. The current collection script is written in Python 3 and relies on Grove’s package functions. <strong>Data collection routines</strong> The Raspberry Pi Zero, which records the current, is always switched on. Every twelve minutes (except for beehive 5, where this frequency varies), it sends a signal through GPIO to wake up the other Raspberry Pi, the beehive data recorder. Once the Raspberry Pi 3b+ wakes up, it collects data from all its sensors, including three 10-second audio samples collected at the same time, five 800×600 pixels images spread over five seconds and temperature and humidity measurements of the inside of a beehive, and it transfers the data through Wi-Fi to a remote data storage cloud server. In addition, at regular intervals, the Raspberry Pi Zero WH transfers the latest consumption data. For now, we share the current data and the temperature/humidity. Later, we will share the audio and the image datasets. The date of the beehive data files from beehives 1, 2 and 3 do not match the current data due to the lack of network during the acquisition process. It prevented the Raspberry Pi 3b+ to update its internal clock during wake-up. <strong>File Structure:</strong> Beehives 1, 2 and 3 are located on the roof of a university (ENS de Lyon) in Lyon, whereas beehives 4 and 5 are located on the roof a university (aivancity Paris-Cachan) in Cachan. Beehives 1 and 4 met technical difficulties which lowered the amount of collected data, whereas beehives 2, 3 and 5 gathered data from April to July 2022 <strong>Folder</strong> <strong>Number of files</strong> beehive1/consumption 6 beehive1/temphumi 37 beehive2/consumption 4158 beehive2/temphumi 2598 beehive3/consumption 3503 beehive3/temphumi 3120 beehive4/consumption 182 beehive4/temphumi 261 beehive5/consumption 1915 beehive5/temphumi 1736 <strong>3- Citation</strong> Works citing this dataset should also refer to [1] as the article that introduces the system used to collect the data. <strong>4- Acknowledgment </strong> The research that led to this dataset was made possible thanks to the funding from aivancity School for Technology, Business &amp; Society Paris Cachan, and the contribution of École Normale Supérieure for the hardware. [1] H. Hadjur, D. Ammar, and L. Lefèvre, “Services Orchestration at the Edge and in the Cloud for Energy-Aware Precision Beekeeping Systems” 5th Workshop on Parallel AI and Systems for the Edge (PAISE 2023), an IPDPS workshop, 2023.