First introduced during NASA’s Apollo missions in the 1970s, the concept of digital twins has come a long way in the past five decades. But what truly put it in the spotlight in recent times was the pandemic. Take the case of contract manufacturer Flex. When a leading global medical equipment company shortlisted Flex to manufacture its Class II diabetes product in the first quarter of 2020, it had a challenge on its hands. The US-headquartered company had to accelerate and optimise the development of the product in the wake of the Covid-19 pandemic while ensuring quality and reliability as any mistakes in the design or manufacturing process could put patients at risk. To optimise the manufacturing of this product, it created a digital twin of the factory floor. With the virtual representation of the complex production system, Flex tested how it would perform under various scenarios and gained insights into optimisation tactics. “Without simulation, this testing process would have taken at least three months longer. But our DES [discrete event simulation] tool allowed us to complete our optimisation analysis in only three weeks, enabling our customer to get the product to market faster. This saved our customer time and money,” says Paul Baldassari, Executive Vice President of Worldwide Operations at Flex. Plus, there were other efficiency gains—a 70 per cent increase in station output by optimising work order sizes, 47 per cent more units produced per quarter, and a 15-21 per cent higher value-added process utilisation.
Back home, Mumbai-headquartered Tata Consultancy Services (TCS) collaborated with non-profit Prayas Help Group to develop a digital twin of Pune city to predict the evolution of the pandemic. “We were able to simulate the likely extent of virus spread as people go about their lives, and the consequent impact on health infrastructure, institutional quarantining infrastructure, testing infrastructure and so on,” says K. Ananth Krishnan, Chief Technology Officer at TCS. This work is being repurposed to help define strategies to effect work-from-home to work-from-office transition, he adds.
But what is a digital twin? In the simplest of terms, it can be understood as a computer program that uses real-world data to create simulations that can predict how a product or process will perform. It can be a real-time virtual replica of a real-world entity of say, an aircraft engine, a factory shop floor or even a human being. “A digital twin is sometimes referred to as a computer-based copy of anything that exists physically,” says Praveen Mysore, Director of Industry Process Consultants and Solutions at Dassault Systèmes India. But it isn’t a standalone technology and involves multiple solutions that include the Industrial Internet of Things (IIoT), connectivity, cloud computing, AI and ML.
With increasing demand from healthcare and pharma industries because of the pandemic, the global digital twin market is projected to grow at a CAGR of 58 per cent to reach $48.2 billion by 2026 from $3.1 billion in 2020, according to MarketsandMarkets, a research firm. While India-specific numbers aren’t available, Mysore believes that because of the government’s push for digitisation and smart city projects, the Asia Pacific region is likely to capture a considerable share, with China, India and Japan leading the way.
Rising popularity
Digital twins have come a long way since inception, and the evolution of technologies such as CAD, IoT, cloud, and big data analytics has made it easier and a lot cheaper to adopt. The initial version was a simple simulation-based event where organisations created the virtual versions of the physical structures to test them for various load cases. The primary advantage was the model’s ability to identify the underlying risks without actually investing in a physical prototype; the virtual model provided unlimited information about the product and its design at various operational scenarios. “Over a period of time, the application of digital twins increased, driven largely by the amount of data that gets generated through instrumented assets. The insights generated through digital twins increase tremendously whenever they can be applied in real-time and invariably brings in an element of predictability in addressing run-time failures,” says Sampath Kumar Venkataswamy, Senior Research Manager for Manufacturing Insights at IDC Asia/Pacific.
Digital twins have gained popularity in India over the past two to three years in spaces such as manufacturing, utilities, and oil and gas, though for most companies it is at an R&D or proof-of-concept (PoC) stage. “In the next three to five years as these PoCs mature, we visualise large-scale adoption of digital twins, targeted towards specific business needs of an organisation,” says a Wipro spokesperson.
Industries such as oil and gas, minerals, metals and mining, power, renewables, pulp and paper, and pharma are focussing on easier optimisation of products and processes by using digital replicas. “For instance, a typical oil refinery can incur losses to the tune of millions due to unscheduled shutdowns and recommissioning efforts. By investing and implementing in a digital twin, these organisations can pre-empt shutdowns and create a targeted ramping down plan to keep production disruptions to a minimum,” says Venkataswamy. Its application can be across the value chain, with organisations looking to understand the supply chain impact and its effect on manufacturing operations, which in turn would affect customer deliverables.
There are several examples of companies using digital twins. Cairn Oil and Gas was one of the first upstream companies to take steps to implement a comprehensive digital operations programme. In India, this private oil and gas explorer deployed Honeywell Forge enterprise performance management (EPM) software to improve efficiency, while enabling workers to remotely operate its facility. Telecom, too, is a big adopter, with Finnish network equipment maker Nokia introducing its 5G Digital Design concept to simulate 5G use cases. Using machine learning algorithms, the platform can leverage this technology to monitor and assess the impact of 5G implementation, while providing automated recommendations. “The solution has benefitted several of our CSP [communications service provider] customers across the globe,” says a Nokia spokesperson.
While an increasing number of automotive companies have adopted digital twins, VE Commercial Vehicles, a joint venture in India between the Volvo Group and Eicher Motors, adopted the 3DEXPERIENCE platform of Dassault Systèmes to develop and deliver innovative, high-quality trucks and buses to the growing commercial vehicles market in the country, in a cost-effective manner. Similarly, Gurugram-headquartered JBM Group—which is into automotive, buses and electric vehicles, among others—uses Dassault’s virtual twin to develop next-gen products, from conceptualising design to manufacturing and even extending to after-sales support. Deepak Thakur, CEO of JBM (Bus Division), says, “Using this technology, we were able to map the entire new product development lifecycle on a single digital thread.”
This technology is also good for the environment. “Carbon digital twins are part of an emerging segment for most companies in India. They can help in visualising the reduction of the carbon footprint of equipment and processes by analysing data,” says Kap Prabhakaran, VP of Engineering at Honeywell Connected Enterprise.
The cost factor
There is no standard cost when it comes to building a digital twin as it varies by company and the sophistication of the technology developed. The cost is largely from the infrastructure needed to generate, store and process the data. Some other key components include digital twin software, IoT sensors, integrated platforms, compute solutions, cloud & infrastructure, and training. Rajesh Gharpure, EVP & Global Head of Manufacturing, Energy & Industry 4.0 at Larsen & Toubro Infotech (LTI) explains: “The initial setup cost for a digital twin is high because of the technology components required to implement a basic digital twin use case like IIoT, connectivity (4G, 5G, 6G), cloud computing, AI & ML and sensorisation of asset/entity, but digital twin is a necessary foundation over which future ‘as-a-service’ business models will be built.”
Typical PoCs or one-plant pilots cost $200,000-500,000, depending on the size or scale of operations and the complexity of the model. Plus, there is the cost of maintaining the digital twin along with the necessary IT infrastructure. The payback period is typically 6-12 months for PoCs with the RoI in certain cases being more than 10x, says a Wipro spokesperson. Take the example of Piramal Glass, which uses Microsoft’s Azure IoT to optimise its manufacturing operations and create a feedback loop between its quality control and production teams. The company is tracking every part of the process from the time the raw material is fed into the furnace to the time its bottles come off conveyor belts. “With Azure Digital Twins, Piramal Glass has achieved 40 per cent reduction in manual data gathering, 25 per cent improvement in employee productivity, and 5 per cent reduction in defects,” says Shivir Chordia, Azure Business Group Lead at Microsoft India. “The RoI is OPEX reduction, improved productivity due to automation, and better run-and-operate models,” explains Nikhil Malhotra, Tech Mahindra’s Global Head of Makers Lab. The IT company had created a complete training landscape for its energy customers in India where a digital twin not only is a digital representation of the landscape but also enables training of company associates around plant operations, enabling OPEX reduction and improved productivity of associates.
For companies, adopting this technology has several advantages like reduced OPEX, improved productivity due to automation, and even discovering errors that are otherwise difficult to find; but the technology can be leveraged to its full potential only when challenges such as availability of quality and real-time data, reliable connectivity and infrastructure, and upskilling and reskilling are addressed.
@nidhisingal