Net Zero and the Digital Transformation
Net Zero and the Digital Transformation

Overview

On the 22nd of April 2016, the European Union, together with over 170 independent states, signed the Paris Accords ― an international treaty on climate change constituting a binding agreement to ‘hold the increase in global temperatures to well below 2°C above pre-industrial levels while pursuing efforts to limit the increase to 1.5°C above pre-industrial levels’.

To meet this challenge, many countries and organisations have committed to achieving carbon neutrality by 2050. Commonly termed ‘Net Zero’, the intention is to fully negate the global emission of greenhouse gases (GHGs) by mobilising a broad array of institutional reforms, including switching from fossil fuels to renewable energy sources, decarbonising the power sector, adopting low carbon , and capturing carbon for safe geological storage.

Given that almost 90% of the world’s primary energy is derived from fossil fuels 1, and that over a fifth of global emissions come from the hard-to-abate sectors of industry and agriculture (the cement industry alone releases more GHGs into the atmosphere than any one country, barring China and the U.S. 2), the challenge of achieving global carbon neutrality within three decades cannot be underestimated.

With so much at stake, the hopes of many now lie with the UN climate summit, COP26, to inaugurate a new level of commitment from governments across the globe and facilitate fundamental changes to the way we live and work, thereby paving the way to a low carbon future.

In addition, over the past few years, another fundamental change has been quietly gathering pace ― one which may play a critical role in the world’s drive to Net Zero: the Digital Transformation.

Based largely on the widespread adoption of artificially intelligent systems, together with significant advances made in the fields of data analysis, communications technology, connectivity and automation, this paper provides an overview of how the Digital Transformation can support the global shift to carbon neutrality.

Digital Transformation Themes

When searching for a definition of Digital Transformation, it’s easy to get lost in a plethora of tech buzzwords and corporate terminology. As digital technology is embedded in almost every aspect of our daily lives (one of the reasons why almost a third of the carbon emissions reduction needed to meet Net Zero could be achieved by leveraging existing technology 3), it comes as no surprise that a single, meaningful definition is hard to find.

However, in terms of the drive to Net Zero, there are certain ‘themes’ within the Digital Transformation that lend themselves to large scale, societal changes. These rapidly evolving technical domains include artificial intelligence and machine learning, advanced connectivity and the Internet of Things, and robotics and automation.

Artificial Intelligence and Machine Learning

In its simplest form, artificial intelligence (AI) and its subfield, machine learning (ML), are established technologies that use advanced algorithms and rich datasets to solve problems, make predictions and perform classifications based on their input data. While the human-like cognitive capabilities of AI/ML can be used to augment almost any process which involves digital technology, here we focus on two key areas where AI/ML enabled processes are particularly powerful: Modelling, Simulation and Forecasting, and Monitoring, Analysis and Control.

Advanced Connectivity and the Internet of Things

Within the auspices of the Digital Transformation, there are many technologies that could be categorised as ‘advanced connectivity’ (for example, 5G, mesh networks, low Earth orbit satellites, fog computing), however, for the purposes of this paper, the term is used to define any developing technology concerned with ‘getting data from where it is collected, to where it is analysed, to where it is needed to drive real-time decisions and automated operations. 4

With advanced connectivity comes vastly improved bandwidths, lower latency and (in a cellular network) an increased capacity for larger numbers of devices per cell ― all of which are fundamental drivers to the emerging phenomenon of the Internet of Things: a vast network of intelligent devices, that can essentially be thought of as an extension of the internet. Spearheaded by the development of smart phones, it is estimated that by 2030, there will be around 50 billion interconnected devices on the planet 5 ― gathering, processing and transmitting so much data that their energy requirements alone will produce more global emissions than the present-day United States. 6

Robotics and Automation

Just as advances in networking and connectivity have turbo-charged the development of IoT, the rise of smart devices and pervasive sensing, monitoring and tracking technology are revolutionising the world of robotics. In fact, the development of IoT and robotics is so closely aligned that the two communities have come together to create the Internet of Robotic Things (IoRT).

Process automation, autonomous technology and the remote management and maintenance of assets are considered to be key components in the drive to Net Zero, all of which are dependent upon intelligent devices monitoring events, processing sensor data and making use of local and distributed intelligence to decide on courses of action and manipulate the physical world.

Emissions Management by Economic Activity

Emissions by Economic Activity

The U.S. Department of Economic and Social Affairs lists over twenty major economic sectors (or sections) containing hundreds of subdivisions 7, however, for the purposes of this paper, the majority of global GHG emissions may be attributed to five economic ‘activities’: Energy, Land Use, Industry, Transportation and Infrastructure.

Of these economic activities, most authorities agree that Energy (including hydrocarbon extraction, refining, processing, transportation and use, plus all the activities associated with the production of electricity and heat) and Land Use (primarily agriculture and forestry) account for almost 60% of global GHG emissions, with industry, transportation and infrastructure making up the remainder.

Advances in digital technology and the themes outlined above may help to reduce emissions associated with all of these activities.

Source: IPCC(2014) https://www.ipcc.ch/report/ar5/wg3/

Energy

Decarbonising the energy sector is primarily concerned with a move away from fossils fuels as a primary energy source in favour of greener renewables (wind, solar, hydro, biomass and geothermal), in conjunction with a shift towards low carbon energy vectors and end-usage ― both predominantly focussed on electricity and hydrogen.

Achieving fundamental changes in the way we produce, transport and use energy comes with countless challenges. However, digital technologies offer opportunities to model, monitor, organise and control future energy systems in support of Net Zero.

Artificial intelligence and machine learning will play a key role in decarbonising the energy sector. For example, one of the greatest challenges in transitioning from a hydrocarbon based energy system to renewable energy (RE) is in maintaining a spinning reserve. This, in essence, is an operating reserve of energy available at short notice to meet surges in demand or interruptions to supply. As many forms of RE are prone to intermittency (the sun doesn’t always shine), energy buffering is currently provided by coal and natural gas power plants. However, to achieve a low carbon world, AI-powered modelling and simulation must fill this role. By accurately predicting energy demand and grid load, combined with forecasting and monitoring short term supply (including the analysis and modelling of meteorological data), energy grids can be balanced in real time. Furthermore, a shift to RE will result in a move away from power plants and towards a highly decentralised power generation network consisting of everything from wind farms, microgrids and solar parks to small-scale run-of-the-river hydro and vehicle-to-grid technology. This complex and highly dispersed network of generators can only be managed by intelligent systems making real time decisions on scheduling and storage based on active monitoring and simulation.

In addition to managing supply and demand, AI/ML based technology can support the development of new materials (e.g. synthetic fuels, HVDC convertors, solar crystallography) and the incorporation of new nuclear energy sources, in particular the implementation of mini reactors. Furthermore, AI can help to maintain existing assets via Digital Twinning whereby a physical asset (e.g., a wind turbine) is digitally modelled to inform asset control loops (see Box 1).

Box 1: Digital Twinning

Digital Twinning

By leveraging both AI and digital connectivity, it is now possible to gather huge amounts of data from physical assets and model them in real time. These ‘Digital Twins’ can then be used to run simulations, predict performance and identify potential faults, while virtual walkthroughs enable skilled personnel to visit sites and perform system checks, no matter where they may be based.

Information and insights derived from simulations and virtual inspections can then be used to schedule downtime and optimise systems in the real world, thereby forming an extremely efficient control loop.

While still in its infancy, it is thought that Digital Twinning will eventually be scaled up to the point of developing planetary control loops, taking into account multiple objectives such as emissions management, biodiversity, resilience and wellbeing 8.

Advances in digital connectivity and IoT may also be leveraged to decarbonise the energy sector. Along with providing real time data to Digital Twins, IoT will be essential in energy system monitoring and management. Furthermore, IoT and robotics can facilitate the transition to remote asset management via component tracking and lifecycle management, workflow automation, predictive maintenance and condition monitoring. Lastly, where direct human intervention cannot be avoided, AI powered IoT will arm the worker with everything from real time data analysis to interactive Head’s Up Displays, supporting on-the-job decision making and providing insights.

Land Use

Accounting for almost a quarter of global GHG emissions, land use is unique in acting as both source and sink in the carbon cycle. However, human mismanagement of land in the form of unsustainable agricultural practices and deforestation are simultaneously increasing emissions while reducing organic sequestration. Furthermore, rice and cattle farming are now responsible for the release of more atmospheric methane than all the world’s natural systems combined 9.

One of the most important practices in which advanced digital technology can help to address this hard-to-abate sector is in the support of precision farming.

In terms of emissions, the tilling of land directly releases carbon into the atmosphere by disturbing soil aggregates and disrupting organisms that contribute to sequestration 10. Further to this, intensive crop growth and rainwater run-off remove much of the soil’s nutrients, necessitating the use of nitrogen based fertilisers which require huge amounts of energy to produce*. Precision farming has been designed to help address these issues.

Precision farming land use

Alongside the analysis and interpretation of soil condition, AI/ML powered technology can support precision farming via weather alerts, intelligent irrigation system management, weed and disease detection, and emissions tracking 11, while intelligent harvesting equipment can provide real time yield data, providing valuable information for fertiliser planning.

IoT and Robotics are also playing a part in land management. Informed via satellite imagery, semi-autonomous drones have already proved themselves useful in reforesting difficult-to-access habitats 12. Meanwhile, tech companies are developing agricultural robots that are capable of picking and packing soft produce, identifying and physically removing pests and weeds, and providing direct sensor data to support decision making.

Industry

Responsible for over 20% of global GHG emissions 13, industrial processes represent some of the most challenging aspects in securing a low carbon future. Manufacturing produces 98% of the total direct CO2 emissions within the industrial sector 14, with the production of steel, cement and chemicals foremost (for example, plastics and fertilisers) ― and most of these emissions result from difficult-to-change chemical reactions and the intense heat which they often require.

The IPCC highlighted six areas where industrial emissions might be mitigated (see Box 2), almost all of which could be facilitated by advances in digital technology.

Box 2: In tackling the difficult-to-abate industrial sector, the IPCC recommends six areas to address:

  1. Improved energy efficiencies. More efficient use of energy to create heat, drive chemical reactions and perform work.
  2. Emissions management: Reducing emissions by limiting the use of fossil fuels. Actively capturing and storing carbon. Improving leakage and waste management.
  3. Material used in production: Material efficiencies gained via process optimisation and switching to lower-carbon materials. Reducing yield losses in production. Re-use of old material.
  4. Material efficiency in design: Less material used in product design. Selecting materials that require less energy to process.
  5. Product use: Using products more intensively. Minimising waste (perishables etc). Producing fewer products while delivering the same service. Using products for longer.
  6. Reducing overall demand: Determined by the end user, industrial emissions would be reduced if people travelled less, heated or cooled buildings more efficiently, and reduced unnecessary consumption of products.

AI/ML can be a powerful tool in materials science and the development of low carbon products. For example, to reduce the use of cement and steel, researchers have combined ML with generative design to develop structural products that require less raw material, thereby much reducing the associated GHG emissions 15. ML-based management systems could also dramatically reduce the power requirements of heating, ventilation and air conditioning (HVAC) systems. Google recently announced that they had managed to reduce the cost of cooling their data centres by 40% using a system of neural networks trained on different operating scenarios and parameters 16.

Another way that artificial intelligence can support industrial emissions reduction is in demand forecasting and supply chain optimisation. Given sufficient data, AI/ML based systems will be capable of predicting supply and demand, identifying lower-carbon products, and optimizing shipping and haulage routes.

Connectivity, IoT and automation will also play a key role in tackling industrial emissions. Aside from the obvious benefits of automating industrial processes in terms of efficiencies, real time monitoring of plant and equipment will enable producers to improve processes, identify and manage leakage, emissions and waste, and highlight potential breakdowns or areas for improvement. Pervasive sensing, tracking and monitoring technology embedded within industrial processes will also help businesses to track and identify equipment for reuse, and provide transparency in carbon accounting.

Lastly, with the continued maturation of 3D printing technology, it is possible that materials and component parts may be produced locally and with lower carbon footprints than traditionally sourced materials 17.

Transportation

Logistically, transportation is one of the most difficult sectors to decarbonize. Of an estimated 1.4 billion vehicles in the world, 99.6% are powered by the internal combustion of fossil fuels, with only 2% of new cars being electric vehicles. Transportation produces the majority of the UK’s emissions, and yet, despite the introduction of increasingly stringent vehicle emissions standards, the UK’s emissions have remained unchanged for a decade as greater fuel efficiency has been offset by miles driven 18.

While battery and fuel cell powered electric vehicles are often cited as the answer to tackling transportation emissions, many authorities believe that much more is required (even if all new cars were fully electric, it would still take 15-20 years to replace the world’s fleet 19). Further to this, ride sharing and vehicle automation (for example, Transport as a Service) may potentially increase GHG emissions due to the Jevons’ effect*.

Despite these challenges, reducing emissions associated with transportation remains a key component in the drive to Net Zero.

reducing emissions associated with transportation remains a key component in the drive to Net Zero

Perhaps one of the most effective strategies in reducing transportation GHGs is in route optimization. AI/ML systems can significantly improve congestion management by augmenting traffic data (via automated vehicle recognition cameras and other remote sensing systems) and modelling its activity, which can then be used to predict congestion on similar roads. Similarly, AI/ML modelling can support freight consolidation and rerouting, thereby helping to manage the complex interaction of shipment sizes, transport modes, origin-destination pairs, special service requirements and congestion avoidance ― all of which will ultimately reduce the amount of traffic on the roads.

Artificial Intelligence can also be used to reduce the energy requirements of vehicles via advanced combustion engine design and vehicle aerodynamics, while ML is already critical to the operation of autonomous vehicles in road and obstacle recognition.

However, alternative fuels and electrification remains the primary means to decarbonise transport, and is utterly reliant on advances in modern technology. From the research and development of batteries, to monitoring and modelling charging patterns and the implementation of vehicle-to-grid technology, AI will be instrumental in the future adoption of EVs.

Lastly, in terms of IoT and Robotics, small autonomous vehicles, such as delivery robots and drones, could reduce the energy consumption of urban or last-mile delivery.

* The Jevons’ effect (or Jevons paradox) refers to a situation where increased efficiency results in higher overall demand. For example, autonomous vehicles could cause people to drive more, so that overall GHG emissions could increase even if each ride is more efficient.

Infrastructure

While transportation represents some of the most ‘difficult to address’ challenges in the drive to Net Zero, many aspects of infrastructure represents some of the easiest. For example, buildings are responsible for a quarter of all energy-related emissions, and yet reductions of up to 90% could be achieved via relatively inexpensive solutions (e.g. adequate insulation, use of heat pumps), many of which could also save costs.

However, retrofitting existing buildings is difficult to standardise and the costs can sometimes take many years to recover. Here, AI-powered analysis can be used to model the energy requirements of individual buildings and predict potential savings, thereby allowing building owners to select the most cost-effective solutions.

In fact, modelling the energy usage of buildings can go much further. AI-based systems, supported by smart devices and sensors, can model energy usage patterns of buildings down to the level of recognising the energy signature of individual appliances, thereby providing invaluable information about usage patterns and efficiencies. This data can then be used in demand forecasting , HVAC system management and behaviour-based optimised heating.

AI-based systems supported by smart sensors and remote sensing data can help to manage the energy requirements of street services while leveraging existing street infrastructure to improve 5G connectivity, monitor air quality and provide vehicle charging hubs 20.

Furthermore, advanced product lifecycle modelling and materials tracking technology could help to identify and locate materials for reuse ― an essential component in developing a low-carbon circular economy (see Box 3).

Box 3: Circular Economies

Circular economy

In a circular economy, material cycles are closed following the example of an ecosystem. There is no such thing as waste, because every residual stream can be used to make a new product. Producers take back their products after use and repair them for a new useful life ― all of which is powered by renewable energy.

In order to function, the circular economy requires a high degree of connectivity. Every actor in the economy (company, person, organism) is connected to other actors, to form a network in which participants become  increasingly interdependent.

Lastly, Artificial Intelligence could play a key role in urban planning. Understanding urban sprawl patterns would help to plan public transport routes, while district heating and lighting depots, power generation facilities and charging point locations could all be better placed and controlled in a hyperconnected, data driven city.

Supporting New Technologies

A Royal Society report on climate change and digital technologies suggested that “the net zero transition should be data-led, with governance arrangements in place that enable the safe and rapid use of data to support the achievement of the net zero target 21”. Encompassing everything from personal carbon accounting to digitally twinning the planet, the ready access to accurate, timely and trustworthy data and information will be a key component to achieving carbon neutrality.

At its heart, the digitalisation of the Net Zero transition means a transition that involves collecting quality data and information on emissions and energy use, and then using that information to create innovative solutions to bring down energy use and emissions.

In order to achieve this, salient data must be easy to find, access and use by automated systems (see Box 4), while communication and collaboration across sectors and industries will require a transparent, common language.

Box 4: The Royal Society recommends using the FAIR principles in the acquisition, use, storage and dissemination of data by automated systems:

Findable: Easily located data and information, including machine-readable metadata.

Accessible: Clear routes to access, including authentication and authorisation.

Interoperable: Easily incorporated into different software and systems.

Reusable: Well described data that can be easily replicated, repurposed or recombined.

Along with the creation of international agreements needed to share data and information, the drive to Net Zero will require in-depth technical collaboration between governments, regulators, industries and the , itself supported by overarching governance structures, data quality standards, taxonomies, search tools and collaboration technology.

Further to this, institutional and personal carbon accountability will require transparent reporting supported by explainable processes, particularly where Artificial Intelligence has played a role. Individuals, as consumers, will need clarity on their carbon footprints which, in turn, must be supported by well-maintained data and audit trails.

Lastly, creating continued widespread buy-in and public support for the drive to Net Zero will require trustworthy data, reporting and technology in order to define the challenges, outline proposed changes and solutions, and report on progress.

Conclusion

Achieving Net Zero emissions by 2050 will require many deep and broad changes to the way we live and work. Every sector will be affected by the transformation to some degree, and many aspects of today’s society will be unrecognisable within a few decades.  Underlying much of this transformation, advanced digital technology will act as an indispensable enabler and catalyst, supporting activities ranging from research into new materials and limitless energy sources, to monitoring and controlling the heating systems in our homes.

There is no doubt that the Digital Transformation will prove to be a significant boon in the drive to Net Zero. However, incorporating advanced digital technologies into our daily routines will present its own challenges. Robust, cross-sector governance models must be developed if carbon accountability and emissions reporting is to be adopted ubiquitously, and technology intended to govern aspects of our lives must be transparent, trustworthy and contestable. In addition, universal data standards should be agreed if planetary-wide systems are to be monitored and controlled, and taxonomies supporting a common language will need to be implemented between all relevant sectors and industries. Furthermore, the sharing of information will play a key role, and as the Royal Society suggests,  “the public, private and third sector will need to devise ways to facilitate working across research domains and collaborations across sectors in order to develop and deploy creative technology solutions. 22

Lastly, In a world where AI-driven, automated systems play an increasingly dominant role, unencumbered access to information regarding those processes will be of critical importance. Here, Intelligent Search tools and advanced reporting technology will prove invaluable, while the adoption of principles ensuring information is findable, accessible and usable will support their use.

Flare Solutions are an active participant in the global digital transformation, spearheading intelligent search and AI-enabled collaboration technology in the energy sector.

Sector

Technical Transformation

AI Modelling, Simulation and Forecasting

AI Monitoring, Analysis and Control

Advanced Connectivity, IoT and Robotics

Energy

Demand forecasting

 

 

Smart Grid real-time optimisation

 

Scheduling and flexible demand of storage

 

 

Batteries and Materials Science

 

 

Waste and Leakage management

Automated / remote asset maintenance

Digital twinning

Demand based energy price modulation

 

Solar and wind facility orientation

 

Energy system integration

 

 

Scope 1 emissions management

Process optimisation in decommissioning

 

 

HVDC conversion unit development

 

 

Synthetic fuel development

 

 

Field development automation

Onsite data processing and analysis

 

Transportation

TaaS and Autonomous Vehicles / Drones

Freight route optimisation and congestion management

 

Vehicle design efficiencies and power management

 

EV optimisation and maintenance

 

 

Passenger transport mode/route optimisation

 

Public transport network operations optimisation

 

Airport logistics optimisation

 

Industry

Materials science

 

 

Low GHG raw material alternatives

 

 

Supply chain optimisation

 

 

Waste and spoilage management

Streamlining systems (HVAC etc) and processes

Digital twinning

Carbon and emissions monitoring accountability

 

Circular economies / material tracking

Remote working and collaboration technology

 

 

Materials science

 

 

Low GHG raw material alternatives

 

 

Land Use

Remote Sensing data augmentation and analysis

 

 

Precision agriculture

Habitat monitoring and maintenance

 

Weather alerts

 

Digital twinning and control loops

 

References

[1] https://ourworldindata.org/renewable-energy
[2] Chatham House Report, Energy Environment and Resources Department, 2018.
[3, 8, 21, 22] The Royal Society: Digital Technology and the Planet, 2020.
[4] Deloitte: Connectivity of Tomorrow, 2019.
[5] https://www.theguardian.com/environment/2017/dec/11/tsunami-of-data-could-consume-fifth-global-electricity-by-2025
[6] https://www.statista.com/statistics/802690/worldwide-connected-devices-by-access-technology/
[7] The International Standard Industrial Classification of All Economic Activities (ISIC) Rev. 4
[9] https://www.giss.nasa.gov/research/features/200409_methane/
[10, 11, 17, 20] Tackling Climate Change with Machine Learning. D Rolnick  et al, 2019.
[12] https://www.forbes.com/sites/jamesconca/2020/09/30/drones-can-reforest-the-planet-faster-than-humans-can
[13, 14] IPCC Fifth Climate Assessment Report, 2018
[15] Rubaiat Habib Kazi et al. Early stage 3D design explorations with sketching and generative design, 2017.
[16] https://www.google.co.uk/about/datacenters/efficiency/
[18] https://www.carbonbrief.org/in-depth-qa-what-is-the-uks-net-zero-plan-for-transport
[19] University of Oxford, True Planet report, 2021.  https://www.research.ox.ac.uk/area/trueplanet
[20] https://www2.deloitte.com/uk/en/pages/annual-report-2021/stories/future-looking-bright.html