The need
Food waste is a huge problem across the globe. The UK alone generates in the region of 9.5 million tonnes of food waste each year. And, with the move to more renewable energy sources such as solar and wind, the cost of fertiliser – a by-product of the oil and gas sector – is set to increase.
In response to such challenges, IntelliDigest is developing a biocatalytic system (iDigest) that uses the precise dosing of enzymes to break down inedible food waste and recover nutrients for use in agriculture, enhancing nutrient dense local food production.
Understanding the composition of the food waste is crucial to optimising the enzyme dose and the efficiency of the iDigest system. IntelliDigest are seeking to achieve this using a combination of low-cost in-line photonic sensors and signal analysis using AI.
The solution
In a project supported by the Measurement for Recovery (M4R) scheme, NPL helped IntelliDigest to make crucial steps towards their goal. Using the facilities of the National Centre of Excellence in Mass Spectrometry Imaging (NiCE-MSI), the NPL Biometrology team collected data using infrared spectroscopy of typical nutrients found in organic mixtures like food waste. These data were then used to train an AI analysis method to predict the concentration of these different nutrients. The trained method proved highly successful when tested blind against mixtures of pure nutrients and examples of food waste.
The impact
For the iDigest system to function optimally, the food waste composition must be reliably detected. AI approaches to analysis have the potential to significantly increase the reliability of signals detected by low-cost photonic sensors. This project provided important proof of principle for the application of AI analysis in the iDigest system and accelerated the product development by 6 months. The intention is that the iDigest system will be used in household and commercial kitchens and will therefore need to be affordable and straightforward to use. The findings from this work have helped IntelliDigest to enable a commercially accessible design.