VLRS - stock.adobe.com
If energy giant Shell were a country, it would be the sixth-largest emitter of greenhouse gases, with the company and its products generating the equivalent of 1.375 billion tonnes of carbon dioxide in 2021, about 2.8% of humanity’s total. By 2050, it has pledged to reduce that number to zero, with software and data playing a significant role.
According to its first energy transition progress report, published in April 2022, Shell has already cut total emissions – including those produced when the oil and gas it sells are burned – by 16% since 2016. Those generated by its own operations are down by 18% over the same period to 68 million tonnes of carbon dioxide equivalent, and the company plans to get to 50% of the 2016 figure by 2030.
Decarbonising one of the world’s biggest fossil-fuel companies means investing in renewables, converting refineries to process low-carbon energy and chemicals, developing carbon capture and storage, and optimising existing processes. “Digital is going to be a very important part of accelerating the energy transition,” says Dan Jeavons, Shell’s vice-president of computational science and digital innovation.
Digital can help make existing processes more efficient; accelerate the design of equipment for processes such as methane cracking, which generates hydrogen for fuel without releasing carbon dioxide; and managing energy systems that rely on variable contributions from sun and wind. A recent analysis by consultancy Accenture for the World Economic Forum said digital technologies could reduce emissions by as much as 20% by 2050 in energy, materials and transport.
But all this requires a common data platform, which in Shell’s case is provided by San Francisco-based software provider Databricks and its data lakehouse model. Much of Shell’s data consists of series of regular measurements. “We have literally millions of these data feeds coming in every day,” says Jeavons, with some 300 million rows of data added every week to 2.7 trillion existing ones.
The data is used for business intelligence on what is currently happening, but it can also be analysed to spot problems at an early stage and improve processes. In Nigeria, Shell has used its real-time process optimiser on settings for equipment within a liquid natural gas plant, which has removed bottlenecks and reduced boil-off gas from evaporation and associated flaring by 70%.
This has the potential to cut carbon dioxide emissions at the plant by 130,000 tonnes a year, equivalent to the total annual greenhouse gas emissions of about 20,000 people in the UK. “Anything we can do on heavy industrial processes to reduce the carbon dioxide associated with production can be very impactful,” says Jeavons.
There is potential to increase the transparency of such operational data so that operators could learn from each other how to reduce emissions: “It lets you compare and contrast,” says Jeavons, adding that it could be used to optimise processes continually rather than on a one-off basis, as at present. But energy companies need common data standards to do this.
“The challenge is that all of them historically have run things on their own software platforms, their own data systems in their own proprietary data formats, by and large,” he says, which has been done in the interests of individual companies. “But if you want to transform the energy system, one of the things we can do is get much more transparent in the datasets we use.”
Shell and Databricks are among those involved in the Open Group OSDU Forum, which is building an open source, standards-based data platform for the industry.
Shell is often targeted by climate change campaigners, and in May 2022, about 50 protestors disrupted its annual meeting in London with chants including “Shell must fall”. But the company argues that it can do more by changing, rather than closing, its businesses. “We’ve made very public commitments to thoroughly transform our business,” says Jeavons. “I can guarantee you that my team every day is working hard to try to figure out how we can use digital technology in support of those goals.”
In acting on data analysis, Shell is currently doing something that only a tiny proportion of users do, according to Databricks’ Junta Nakai. Another example is UK-based aerospace and defence group Rolls-Royce, which estimated last year that it had saved its airline customers 200,000 tonnes of carbon dioxide emissions since 2014 by improving efficiency through its engine conditioning monitoring service and the predictive maintenance this supports.
Dan Jeavons, Shell
Many others are likely to follow, as is illustrated in Nakai’s job title of vice-president for sustainability and financial services, with the former added as a result of the growth of environmentally focused financial services. This means companies with a good environmental record or strong plans may find it easier and cheaper to raise funding, but to make such judgements, investors need reliable data.
Nakai says that one customer, a large pension fund, realised that the ESG (environmental, social and governance) ratings it received from different suppliers barely correlated with each other, so it now uses Databricks to do its own. “It can now take a very quantitative approach to how green or ethical an investment is,” he says. US-based ratings agency S&P Global uses the supplier to provide customers with an ESG investment analytics service.
There is also increasing pressure from regulators. In March, the US Securities and Exchange Commission (SEC) published proposals to require companies that it regulates to disclose their emissions from their own operations (known as scope 1) and those from energy suppliers (scope 2), as well as those from their other suppliers and customers (scope 3) if they have a target for this. “Today, it is treated almost as a second-class citizen disclosure, because it’s not as material as revenue or earnings,” Nakai says of emissions data.
The SEC proposals would change that, meaning that more effort will have to put into collecting and checking emissions data, and could also see it move from annual to quarterly or monthly reporting.
Evaluating scope 1 and 2 emissions can be challenging with older equipment that was not designed for digital processes. However, such components can be monitored with internet of things (IoT) devices, such as optical meter readers that can read water, electricity and gas meters and industrial gauges.
Read more about data analytics and sustainability
- Carbon emissions data to become key factor in cloud purchases by 2025, predicts Gartner.
- CIOs need to balance tech with business sustainability.
- The global state of datacentre sustainability.
Devin Yaung, senior vice-president of group enterprise IoT products and services at Japan-headquartered technology group NTT, says adding such devices can be easier than installing smart meters. His group offers a range of IoT services for sustainability, also including sensors for water leak management, predictive maintenance, pollutants, temperature and humidity.
Outside industrial environments, IoT monitoring can make buildings more efficient, such as by using occupancy monitoring to decide whether to heat, cool or clean rooms. Many organisations need to start by establishing their current positions, says Yaung. “What a lot of companies lack are baselines – how efficient are you today?” he adds. “Most people are running blind.”
Scope 3 emissions from suppliers are harder to track than scope 1 and 2, but some are getting more accurate. Washington state-headquartered travel and expense manager SAP Concur offers data from London-based emissions intelligence service Thrust Carbon that covers emissions for all types of business travel and hotels. Although such data is not exact, for flights, Thrust Carbon uses mileage, aircraft type and average passenger numbers for its estimates. “It is continuing to become more sophisticated,” says Ami Taylor, senior director for product strategy at SAP Concur.
Taylor says organisations now use four Cs – cost, convenience, carbon and care – in setting guidance or rules for business travel and can use data to estimate the impact of, for example, increasing the minimum flight length where business-class travel is allowed. In Europe, the most obvious change is to move journeys from plane to train, which typically cuts emissions by 80%. One customer, a large pharmaceutical company, has shifted 65% of flights to rail.
Other journeys can be replaced with digital meetings, but Taylor says some things work best face-to-face, particularly for staff who have never met their managers or collaborated with colleagues because of Covid-19. SAP Concur research involving 700 European corporate travel executives found that about four-fifths of their organisations have returned to pre-pandemic levels of domestic travel and three-fifths for international travel. But they want to improve sustainability.
“People are thinking about travel in a more purposeful and more thoughtful way, travelling when it makes sense,” says Taylor. One way to do this is by organising longer events that get more out of the travelling required by combining a set of activities for staff.
In some sectors, suppliers are offering tailored carbon measuring services, such as Berlin-based Vaayu for retailing. It uses data from point-of-sale systems and estimates emissions based on three main areas – products, packaging and delivery. Chief executive and co-founder Namrata Sandhu says that reliability of data varies, so the system provides an accuracy level for every transaction. Basing the data on actual transactions means it is available quickly, she says, adding: “You can start to track your carbon emissions on a daily basis.”
Some emissions can be cut with little or no impact on shoppers, such as changes to secondary packaging used in transporting goods to shops, or redesigning products designed for sale in shops for home delivery. Also, customers may be asked to act differently, such as by informing them of the emissions that would be caused by returning an item. This data can be used to demonstrate retailers’ environmental credentials, with London-based jewellery maker Missoma publishing its delivery carbon emissions through Vaayu, showing that in mid-June, it averaged 3.72kg of carbon dioxide equivalent per shipment over the previous 30 days.
Ami Taylor, SAP Concur
A few organisations go even further, regularly publishing a range of data. The City of Reno authority in Nevada discloses its monthly scope 1 and scope 2 emissions and its daily power usage, including the proportions of each hour’s power from solar panels on its own buildings. The service is run by nZero, which is headquartered in the city. Until March, it was named Ledger8760, the number of hours in a year, and among other things, it encourages users to gather and pass on electricity use automatically based on hourly or shorter increments.
Electricity’s carbon emissions can vary dramatically across the day, even for those who rely entirely on grid supplies. Josh Weber, co-founder and executive chairman of nZero, says one customer operating a large building for hospitality found it was under-reporting its emissions by more than 200 tonnes of carbon dioxide a year because it used more electricity in the evening. In London on 6 June, the company calculated that fully charging a Tesla Model 3 would have produced 13.7kg of carbon dioxide at about 10am, but 17.7kg at about 10pm. Grid electricity’s carbon intensity can vary much more in other locations, depending on usage patterns and local renewable generation.
Weber says many clients pay for nZero’s work from marketing budgets, as organisations seek to show they are working to reduce emissions. But detailed data can be used to work out carbon return on investment, leading one nZero client to decide that replacing gas heating with an electrical powered heat pump was a better bet than installing solar panels.
Although perfect data on emissions data is not possible, it can usually be greatly improved, says Weber, adding: “We may not have all the data, but we know we can access better data than what is standardly used by companies and organisations today. You have got to have data to make data-driven decisions.”