A key part of what we do involves getting more from less. This quest for increased efficiency is a cornerstone of the CEA’s drive to deliver a sustainable solution to growing. The industry is constantly after new ways to minimise resource use while maximising crop growth. For us at Zero Carbon Farms, such innovation manifests itself via the development of the first instance of a digital twin for urban-integrated agriculture. In other words, a digital twin of our farms.
Digital twins are more than just a 3D representation of the real world. A key digital trend in the UK, their actual definition is still being debated, as they can have various degrees of complexity. However, they are generally recognised as a virtual representation of existing environments, fed by streams of real-time data, and which can provide feedback. Importantly, they display information that would otherwise be missed. Users are therefore able to use digital twins to make informed improvements to the monitored environment.
For the past 15 years, digital twin technology has slowly become embedded in an increasingly wider range of industries and activities, opening new avenues within industry innovation and smart manufacturing – even Bradford city centre is getting a digital twin!
Since 2017, in collaboration with Cambridge University and the Alan Turing Institute, we have been developing a digital twin of our own. While the former have helped us acquire and analyse the data, the latter are assisting in the final software development of the twin and are currently working on the next iteration; integrating the models developed to provide real-time feedback.
The development and use of a digital twin within our farm is a pivotal milestone for us and, we believe, sets us apart from our competitors.
Throughout our growing environment, sensors have been strategically placed to monitor more than 80 variables. The sensors used communicate the information they record every 5 minutes to the farm’s virtual sibling in a laboratory in Cambridge. This data is subsequently analysed alongside the physical properties and conditions of our underground growing environment to generate models of the farm.
The digital twin is constantly developing, and in time, it will have the capability to act as a 24/7 smart companion to our growers, informing them of real-time conditions in our farm and providing them with solutions to any problems which may arise.
For example, one such problem we encountered was with temperature control. Regarding this issue, the first iteration of the twin helped our growers see how hot the farm gets, where it gets hot and crucially informed them of why it does so. This led to the adjustment of the lighting schedule so that the LED lights came on one hour later in summer compared to winter. The twin provided this extra level of insight you could not get from just walking through the farm.
Ultimately, the development of a digital twin for our growing space will help our growers tweak conditions – such as tailoring light and heat use – to maximise the taste, nutrients and yields of our crops, all while reducing the amount of inputs required for this such as water usage, energy usage and crucially, time. Gradually, such incremental adjustments combine to make a meaningful difference in our results. Already, in some cases we have halved the number of growing days while simultaneously improving yields by around 20%.
But why stop there? Dr Melanie Jans-Singh, our current in-house Data Scientist, took the lead on building the current version of our twin as part of her PhD research on the integration of nature within existing infrastructures, and is helping us with the next stage of both the twin and our company’s development.
Further models based on the existing twin will enable us to understand the requirements of new and diverse spaces for operating from. With this, we can take any empty site and use our data and models from the past five years to work out exactly what is needed to turn that space into a functioning controlled environment farm. This is exactly what we will do at our next site which is scheduled to come online in November 2021.