In March last year, it was announced that digital twin technology had been used in the Phase One development of the toll-free Pan Borneo Highway, which aims to connect both Sabah and Sarawak with Brunei and the Kalimantan region in Indonesia.
Using integrated technology provided by US-based engineering software company Bentley Systems as well as several other third-party solutions, the project’s delivery partner Lebuhraya Borneo Utara Sdn Bhd (LBU) has managed to capture aerial images of the highway and build accurate 3D models of the project.
By integrating the technology with several other systems, such as road information, bridge management and maintenance management systems, a digital twin solution was created to manage the data collected and visualise it through a dashboard for the Malaysian government to monitor and ensure that LBU meets its commitments on the project.
Three years ago, research and advisory company Gartner named digital twin technology as one of the top 10 technological trends of 2017. Last year, Japanese telecommunications giant NTT identified it as one of the top technological trends of the future, but applications of this technology in Malaysia are still thin on the ground.
In fact, the Pan Borneo Highway is the only widely publicised case study of digital twin technology being used by the public sector in Malaysia at the time of writing. A quick search on Google Trends finds that the technology has not received the same level of attention and publicity in Malaysia as other IR4.0 initiatives such as artificial intelligence (AI) and the Internet of Things (IoT).
Yet, the technology has been widely used and adopted by major companies worldwide. Well-known digital twin solution vendors include Microsoft, Siemens, General Electric, Cisco, IBM and Bosch. According to Deloitte’s Tech Trends 2020 report, the global digital twin solution market was worth US$3.8 billion in 2019 and is projected to reach US$35.8 billion in value by 2025, representing a compound annual growth rate of more than 45%.
Digital Edge spoke to experts to better understand the adoption trend of digital twin technology in the local market, and the opportunities and challenges presented to companies looking to implement this technology.
What is digital twin technology?
Put simply, a digital twin is a virtual model or representation of a physical process, product or service, formed through the generation or collection of data — a process similar to the ideas presented in the 1999 science-fiction movie The Matrix.
By pairing both the virtual and physical worlds, users are able to detect problems during the manufacturing process before they even occur, minimise or prevent equipment downtime and even conduct virtual simulations without risking the physical process.
According to a 2019 trend report published by logistics company DHL, global organisations from a wide array of sectors ranging from manufacturing and logistics to life sciences and e-commerce have been developing, testing and using digital twin technology in their operations. This is due to the technology’s innate ability to represent any physical object, from nanomaterials to entire cities, or even people’s behavioural patterns.
Yokogawa Engineering Asia senior vice-president Dr Darius Ngo tells Digital Edge that digital twin technology is not exactly a new innovation but has been in the market for decades.
“Fundamentally, digital twins started with simulation technology, which has existed for a long time, especially in the oil and gas industry. Applications for petrochemicals are quite mature, with simulation libraries since the 1970s,” says Ngo, who also heads the company’s digital enterprise segment in Singapore.
“Digital twins have recently gained attention because of the rising trend and integration of many other technologies that enable digital twins, such as AI machine learning, software analytics and the utilisation of 3D graphics.
“This is also partially because gathering data from physical sites has become cheaper. IoT sensors used to be quite expensive but are reasonably priced these days, combined with the fact that leveraging cloud-based solutions and applications has become possible and easier.”
Ryan Sim, a partner at the consulting arm of Ernst & Young Advisory Services Sdn Bhd (EY), tells Digital Edge that the technology had its beginnings in highly sophisticated and high-risk industries, where real-world testing is extremely costly and the consequence of failure is grave, such as the case of rocket launches by NASA.
“Digital twins can be broken down into many types. The first is the process twin, which can simulate the effectiveness of the logistics process in the production line or supply chain, where data touch points can start from the factory to the ports and, finally, to the customer,” says Sim.
“Next is the component twin, where engineers design a new component to reach a certain optimal output, such as rotary blades for aerodynamics. Finally, there is the asset twin, where you gather data points on the assets or machinery, such as the sound or rumblings, and pair them with the environmental parameters to predict how the asset is going to perform.”
Digital twin market in Malaysia
According to Ngo, Asia is rapidly gaining awareness of digital twin technology but technological adoption is still in the preliminary stages, with many companies still conducting surveys or testing out proofs of concept.
“As for the Malaysian market, Yokogawa has engaged and held discussions with several local manufacturing company owners. They are interested in adopting this kind of solution, but many of them are concerned about their return on investment and how fast they can achieve it,” he says.
“Especially during this current pandemic, many of them are worried about whether they should lay off some of their staff after implementing these solutions. Actually, they do not need to do so. In fact, they may need to hire more people to focus on the specifics within the complex supply chain.”
EY’s Sim believes small and medium manufacturing companies also understand the underlying concept of digital twins, without knowing the actual term.
“The manufacturing sector has been using smart machinery in its production line without awareness of the concept in its entirety. Data is being digested from these devices as part of the work scheduling and maintenance, which is effectively what a digital twin solution is all about.
“More commonly, their use case for digital twin technology is much more targeted. For example, a factory owner would just like to know when a machine will break down and have the system send a notification to the engineer to fix it. These types of solutions already exist in the form of smart sensors and IoT devices.”
He further points out that if the definition of digital twin technology were not so loose, with data being collected in real time and simulations being conducted, the number of actual use cases in Malaysia would be few and far between.
Sim says digital twin technology, however, has become an umbrella term that encompasses solutions that vary wildly in structure and use cases and, ultimately, it depends on how the companies define the term and how the technology is used to benefit them.
Implementing digital twins
Although digital twin technology revolves around the usage of technical devices and complex data, Ngo points out that the biggest challenge in implementing the technology is not technical but, instead, in reshaping the company’s business processes and human resources to allow the technology to be useful.
“When clients wish to embark on this digital twin programme, we have to look at four main factors — practices, systems, people and assets. The devices themselves are only part of the assets, and the challenge revolves around how the organisation makes use of the technology,” says Ngo.
“For example, the client may wish to assign a team to lead the implementation of a digital twin system, but the business process may not even require a digital twin system to be effective. There are also other aspects to consider, such as the personnel’s job description and competency model.”
A common misconception that clients have, Ngo points out, is that vendors are able to implement a digital twin solution without much input from the client. He says it will be almost impossible to do so.
“To build an accurate model of the plant, we need to have the insights on the actual plant itself. So, most of the time, we still need the client’s specialist to be part of the team so that they will also know how to manage the system,” he says.
“Digital twin solutions are not like an app that you can download on your phone and expect to run smoothly. Although we do have a generic principle model for clients to start with, it needs to be tweaked to the client’s usage because the operation methodology may differ from plant to plant.”
In terms of pricing and feasibility, Ngo says digital twin technology is actually accessible to medium-size companies, and its implementation is mainly determined by the level of fidelity the client wishes to incorporate in its digital twin model.
For example, a generic principle model can be used to provide a general overview of the plant’s operations. To derive the benefits from the heart of the plant’s operation, however, much more sophisticated software has to be used to address the company’s specific issue.
According to EY’s Sim, many manufacturing companies in Malaysia are uncertain about taking the first step in implementing the digital twin solution within the company.
He says these companies plan excessively and require a clear picture of the final outcome before committing the resources to do so. He believes, however, that technological advancement is moving at such a fast rate that companies will find themselves waiting perpetually for the “next best thing” if they wish to maximise their return on investment.
Sim explains that it is unnecessary for small and medium companies to fully commit to a highly sophisticated digital twin solution, as most of them would benefit from having a few IoT devices and a simple database.
“There are even companies that offer Digital-Twin-as-a-Service (DTaaS). This means that you do not need to worry about setting up the infrastructure, hiring data scientists or even installing the simulation software, as you can have professionals to manage these for you,” he says.
“By outsourcing these tasks, small companies do not need to invest much capital upfront to achieve a similar outcome.”
Digital twin use cases
Logistics company DHL has been researching and placing emphasis on the implementation of digital twin technology in the logistics industry for years. The result is a partnership with Swedish-Swiss food packaging company Tetra Pak, and the 2019 launch of DHL’s smart warehouse in Singapore, the company’s first in Asia-Pacific that also deploys the technology.
The physical warehouse is bridged with a virtual representation that monitors and simulates both the physical state and operations of its assets in real time, enabling the warehouse to coordinate its operations 24/7 and resolve any issues that might arise, particularly those revolving around safety and productivity.
According to company documents, the warehouse supervisors use real-time operational data to make informed decisions to reduce congestion, improve resource planning and allocate workload. Internet of Things (IoT) and proximity sensors are placed on material handling equipment (MHE), thus reducing potential collision risks. Controlled areas with restricted access are also monitored with management alerts.
A control tower monitors the flow of inbound and outbound goods to maintain time efficiency, ensuring that goods are correctly shelved within 30 minutes of receipt and delivery-bound goods are ready for shipment within 95 minutes.
Justin Baird, the head of DHL’s Asia-Pacific innovation centre, tells Digital Edge that digital twin technology is still emerging and there is some distance to go before a fully realised digital twin system can be built. However, he has already seen how the additional layer of data visibility has increased the company’s productivity and improved safety levels within the warehouse’s operations.
“The company looks into trends that could affect and transform the logistics industry, and digital twin technology first appeared on our radar as a potentially transformative technology that was closely related to developments in virtual reality,” he says.
“We see the rise in importance of digital twins as being driven primarily by parallel advancements in IoT, big data, AI (artificial intelligence), cloud computing and digital reality. Digital twins are a logical confluence of these trends, combining these complementary technologies and creating mirror-image digital copies of objects in the physical world.”
According to Yokogawa Engineering Asia senior vice-president Dr Darius Ngo, the company has been providing digital twin solutions and consulting services for more than seven years, although this technology has become popular only recently.
One of its past clients from the oil and gas industry approached it with an issue regarding the company’s hydrocracking unit, which cracks the heavy molecules in gas oil to produce gasoline.
Apparently, it takes three days for the hydrocracker unit to restart if it unexpectedly shuts down, which resulted in a 60% drop in total production output for the client, who lost an estimated US$2.8 million for each shutdown. The team managed to implement a digital twin system to monitor the unit’s feed flow using IoT devices that issue a warning if there is a problem so the operator can rectify it before it results in a shutdown.
In other projects, Yokogawa has also implemented solutions that can predict an equipment’s tendency for failure through its vibrations, as well as determine pipe corrosion levels using AI machine learning algorithms.