Decarbonisation Technology - November 2022

transportation of a product for one company is also transportation of material for another company • Fair attribution of total supply chains Given that the total GHG emissions for a supply chain have been successfully counted, what is the fair responsibility of each actor in the supply chain? AI-based tools can help establish baseline Scope 3 emissions for companies as they are used to model an entire supply chain. The tools can quickly and efficiently sort through large volumes of data collected from relevant sensors. If a company deploys enough sensors across the whole area of operations, it can identify sources of emissions and even detect methane plumes (Seeq, 2022). Digital twin optimisation A digital twin is an AI model that works as a digital representation of a physical piece of equipment or an entire system. A digital twin can help the industry optimise energy management by using the AI surrogate models to better monitor and distribute energy resources and provide forecasts to allow for better preparation. A digital twin will optimise many sources of data and bring them onto a dashboard so that users can visualise it in real-time. For example, a case study at the Nanyang Technological University used digital twins across 200 campus buildings over five years and managed to save 31% in energy and 9,600 tCO₂e. The research used IES’s ICL technology to plan, operate, and manage campus facilities to minimise energy consumption (IES, 2022). Digital twins can also be used as virtual replicas of building systems, industrial processes, vehicles, and many other opportunities. This virtual environment offers much more testing and iteration, so that everything can be optimised towards achieving its best performance. This means digital twins can be used to optimise building management, creating smart strategies that are based on carbon reduction. Predictive maintenance Predictive maintenance of machines and equipment used in industry is now becoming common practice because it saves companies

costs in performing scheduled maintenance or fixing broken equipment. The AI-based tool uses machine learning to learn how historical sensor data maps to historical maintenance records. Once a machine learning algorithm is trained using historical data, it can successfully predict when maintenance is required based on live sensor readings in a plant. Predictive maintenance accurately models the wear and tear of machinery that is currently in use. The best part of predictive maintenance is that it does not require additional costs for extra monitoring. Algorithms have been created that provide accurate predictions based on operational telemetry data that is already available. Predictive maintenance combined with other AI-based methods, such as maintenance time estimation and maintenance task scheduling, can be used to create an optimal maintenance workflow for industrial processes. Conversely, improving current maintenance regimes, which often contribute to unplanned downtime, quality defects, and accidents, is appealing to everybody. An optimal maintenance schedule produced from predictive maintenance prevents work that is often not required. Carbon savings will be made via controlling deployment of spare parts, less travel to the site, and less hot shooting of spare parts. Intervening with maintenance only when required and not a moment too late will save on the use of electricity, efficiency (by preventing declining performance), and human labour. Additionally, systems can employ predictive maintenance on pipes that are liable to spring leaks, to minimise the direct release of GHGs such as HFCs and natural gas. Thus, it has huge potential for carbon savings. Research has shown that underpinning the scheduling of maintenance activities on predictive maintenance and maintenance time estimation can produce optimal maintenance scheduling (Yeardley, Ejeh, Allen, Brown, & Cordiner, 2021). The work optimised the scheduling by minimising costs based on plant layout, downtime, and labour constraints. However, scheduling can also be planned by optimising the schedule concerning carbon emissions. In this situation, maintenance activities can be performed so fewer journeys are made and GHG emissions are saved.