Our project extension proposes to create and include an AI-vision recognition for trees into the existing CarbonCultivator software. This would enable users to see the project through from initial assessment, to verification from remote imaging. CarbonCultivator would thereby provide a tree-species recognition platform and trained dataset that would assess and verify growth of projects at scale. The programme would also enable the application of species-specific carbon sequestration rates.

Our project aims to reduce the administrative and bureaucracy for farmers and land owners in agriculture to enter the market for voluntary carbon credits. Our project will look to assess the feasibility of using AI to minimise the amount of testing required and to understand what the capacity for carbon sequestration could be for a section of land and to monitor carbon sequestration levels. This will provide land owners with a clearer value proposition for woodland planting and an independent and reliable verification of carbon sequestration. The framework for this project could ultimately be used to assess land in a process similar to an EPC for a property, considering the existing carbon sequestration state of land, and it's potential, supported by recommendations for improvement.

AutoEPC is a commercially scalable solution that uses self-learning algorithms to provide accurate fabric performance with no specialist equipment, installation or monitoring. According to the European Commission, household energy consumption accounts for roughly 20% of total CO2 emissions in UK and Europe, and heating losses are responsible for half of this figure. It is therefore essential to understand buildings' fabric performance in order to reduce their environmental footprint and incentivise projects that align with net zero goals. Existing EPC certificates are too inaccurate due to manual input and calculations, whilst expensive sensors setups combined with long term controlled testing aren't cost effective and therefore scalable across domestic and industrial markets. This project address the technical challenges for large scale commercial adoption, focusing on automation of model setup, data efficient robust to noisy and limited data and exploring the opportunities good quantification of fabric performance in a home can have in realising carbon crediting or incentive opportunities of domestic and commercial buildings.

**Upskilling Future Flight Engineers and Analysts** We are building a 3-month, part-time course which will upskill junior Future Flight engineers and analysts in their workplace. It's called "AI in the Wild: Foundations in Machine Learning for Future Flight". These engineers/analysts typically possess domain-specific knowledge about their systems and familiarity with their data. But the national AI/ML skills gap means that over 90% of data isn't even being used, and valuable insights are being missed. digiLab wants to change this. **Course Structure** Each month will include three weeks of lectures, practice exercises, and seminars; in the fourth week, we will deliver an in-person "hackathon" skills challenge, so that managers can see learning being put into action. We will also be keen to identify opportunities for further collaboration on Future Flight projects. Month 1 will teach Key Principles in ML; Month 2, Deep Learning; and Month 3, Probabilistic ML. **Learning Outcomes** It's called "AI in the Wild" for two reasons. Firstly, we want to empower learners to be able to tackle any problem independently, "in the wild': this will involve teaching not just how to use ML tools, but also how to approach new, unexpected, and varied data challenges. Secondly, through our extensive experience consulting for organisations operating in Future Flight sectors, we know that much of the data coming out of safety-critical environments or physically complex systems is "wild" - that is, messy or sparse. This could be due to malfunctioning sensors or an inability to acquire sufficient samples. So we will be teaching engineers how to reduce computational cost and increase confidence in their model outputs. **Upskilling Benefits** Upskilling employees increases productivity, retention, and wellbeing. Our course will accelerate business outcomes in Future Flight organisations and reduce their dependency on expensive external consultants to plug the in-house skills gap. Ultimately, the UK will benefit from the closing of the AI/ML skills gap and the acceleration of the Future Flight programme through the intelligent use of data.

Solar energy is a clean renewable energy source, however it comes with significant challenges to deploy at scale. This is because energy production does not always align with energy demand, and excessive energy cannot either be easily fed back into the larger electric grid, or the viability of many energy storage solutions is not clear due to losses at conversion or cost. Solving this challenge is central to meeting UK's commitment to generating all our power renewably by 2035\. In this project digiLab will develop a novel digital pipeline to accurately and efficiently optimize the design of urban powerplants using physics informed digital twins. The solution will use self -learning algorithms to optimal balance, production / demand and storage. The innovative approach provide easy to train and fast deployable solutions not currently available in the market, using bleeding edge methods in so-called "Physics Informed digital twins". Whilst the deployment of urban solar production and intelligent control has broad application across the UK and internationally. In this Fast Start project we will focus on a 30MW solar farm on the roof tops of Marsh Barton Industrial Estate, on the edge of Exeter. UPS is looking to generate sufficient power from local solar to take 150 electric delivery vehicles off grid. The opportunity for the particular case study considered at Marsh Baton is clear. By electrifying 150 vehicles with renewable solar, it is equivalent to the equivalent carbon absorption of 200,000 trees annually. Yet such a solution is not possible without our intelligent control algorithm living beneath. This project paves the way to using smart digital technologies to maximise the economic viability of locally balanced energy solution - a necessary step in our net zero journey.

Next Wing will develop a series of Scalable Models for Aircraft Robust Trades (SMART) models and enabling modelling and simulation capabilities, which will be integrated and subsequently demonstrated at wing level. These are critical enablers to shorten the aircraft Product development cycle and develop products that will deliver on Airbus' sustainability ambitions. The project will deliver: * Wing-level SMART models for components and zonal integration areas * Development of Model Based Systems Engineering-based design environment for SMART model integration * Integration and V&V of models in the design environment via use cases * Development of a collaborative co-development approach * Development of a wing Product line approach The partners collaborating in the delivery of this ambitious project are: Airbus Operations Ltd, Capgemini UK Plc, Daptablade Ltd, Imperial College London, Loughborough University, Queen Mary University London, University of Exeter, University of Manchester, and University of Sheffield.