Refuelling locally-produced renewable hydrogen in a tractor is a HyPErFarm goal. Hydrogen panels, developed by KUL, will produce hydrogen from sunlight and ambient humidity and be connected to a small refuelling station (compression, storage and dispensing). We defined technical requirements for the dispenser so hydrogen gas can be refuelled on-farm in a safe and cost-efficient way. Common working pressures for hydrogen refuelling are 350 and 700 bar. Refuelling at 700 bar requires cooling and thus higher investment and operational costs. 350 bar is the choice for a small scale station.

Sensors/components that should be integrated:

- A tilt switch sensor recognizes when the position of the dispenser is skewed.

- Emergency stop

- A hydrogen detector can detect concentrations 10x lower than the explosion/fire limit, before a flammable mixture occurs.

- Hydrogen flames are invisible: a flame detector measuring temperature of infrared radiation is needed.

- Driving away with the refuelling hose still attached, is the main safety accident in stations. A break-away coupling releases the hose with no risk of a hydrogen leak.

- A pressure relieve valve releases hydrogen when the pressure in the system threatens to become too high due to rise in ambient temperature or malfunction of the compressor.

Other components are pressure and temperature sensors and a controlled valve or orifice to manage the mass flow of the hydrogen. A valve is expensive (10 k€), therefore a dispenser with an orifice has been chosen in the project.

It was examined that assembling the dispenser ourselves did not lead to a cheaper one. Therefore, a commercial dispenser, with CE certification and all the required safety equipment, was selected for the HyPErFarm project.

The idea of agrivoltaics is still unknown by many people, and such an innovative approach is therefore in the development process.

To support a customer-oriented view in the development process of agrivoltaics, we used "personas". Personas are known from user experience design, where they represent a fictive consumer. Personas are crafted based on different information, such as interviews, workshops, or consumer behavior statistics. By crafting personas, one designs one or more fictive characters that represent typical customer characteristics of someone who would need or like to buy the product in the future.

In the case of agrivoltaics, we designed different personas that represent the most typical stakeholders and their motivations, needs, goals, and frustrations concerning agrivoltaics. These personas can support those who develop agrivoltaics to avoid failure in the product development process.

Our personas are meant to be of high practical support for those who want to understand the different stakeholders' views on agrivoltaics and those who wish to use a customer-oriented approach when designing and offering agrivoltaic installations. Last but not least, they can be a valuable input for policymakers that decide if and how agrivoltaics might become an approach to contribute to more sustainable farming.

Qualitative data can be collected through interviews or focus groups. Such data is often a lot of text: Conversations are recorded, noted down, and prepared for analysis.

Qualitative research is chosen when we do not know much about a specific topic in advance. We do not simply want to ask if somebody likes or dislikes an innovation. Instead, we want to learn and understand the reasons behind it. We chose qualitative interviews to gain rich and meaningful information, especially in the case of an innovation that is not known by many, such as agrivoltaics. A simple "yes or no," as often answered in a survey, does not tell about a person's thoughts or feelings. Therefore, we openly asked "how and why" questions: How do you perceive agrivoltaics? Why do you think dual land use can be a good [or a bad] idea?

We purposefully decided on who to interview. To better understand the perception of agrivoltaics, we talked with farmers, politicians, journalists, researchers, farmer and nature organizations, municipalities, or photovoltaic producers. Due to the pandemic, all interviews were conducted online. The online interviews were perceived as participant-friendly because of the flexibility to arrange them because participants felt more relaxed in their familiar environment, and it allowed us to interview in Belgium, Germany, and Denmark at almost no costs.

Due to this intensive approach, one can not talk to hundreds of people simply due to time and financial restrictions. We stopped after 27 interviews when no new topics arose. Therefore qualitative information is not claimed to be representative of the whole population. Instead, qualitative research offers intense information and detailed insights into how key stakeholders perceive agrivoltaics.