For decades, product, process and engineering have been improved by using 3D rendering of computer-aided design (CAD) models, asset models and process simulations to validate manufacturability and increase efficiency. However, the convergence of several technological innovations has brought about a new phase in the digital representation of physical objects, with an increasing range of possible applications.
The development and application of digital twins has accelerated in recent years. Digital twins are cloud-based virtual representations of physicals assets, hence have profited from the decreased cost of cloud computing and data management. The decreasing costs in nanotechnology – particularly sensor technology – and increased digitization of our living world has made it easier to feed digital twins with data from our devices and things we do. By connecting digital twins with physical assets using real-time data, physical products and processes are virtually reproduced into a digital environment. Here, machine-learning algorithms further analyze these physical assets, thereby reducing downtime and maintenance costs, R&D costs by digital prototyping and analyze further efficiency improvements into the ecosystems.Although virtually representing physical objects and processes is not new, digital twins are another evolution of the digitization of our living world. This starts at the most basic level, as particular aspects of physical products and processes are simply mirrored by digital data systems, e.g. rows of data about the specifications of a car, an Excel with the groceries we buy or our mobile browsing history. This information can is used to physically calibrate and test the product or process to a set of preferred values and metrics, such as reducing leakage within car motors or boosting online grocery purchases. The next phase is digitizing the whole product or process lifecycle, and feed it with real-time data. In the context of the Internet of Things, these digital datasets are used to improve product lifecycles and in the Fourth Industrial Revolution to improve and enhance production processes. Cloud-based digital twins are another step in the evolution in the digitization of our living worlds, albeit with revolutionary implications. First of all, multiple and different cloud-based datasets about physical products and processes can be connected, hence fully simulating ecosystems, such as life in a biological ecosystem like a pool or a factory’s entire throughput process. This core aspect of removing the silos by merging metadata from different sources and processes, makes digital twins suited for predictive analysis of products and analyzing interactions and impacts on complex systems. Digital twins combined with real-time data will boost predictive repair for products, such as cars that get a message that they are up for maintenance or patients that need to see a doctor because their personalized digital health record shows anomalies. This means that consumers and business will perceive their products as permanent, semi-finished products, which will increasingly stimulate ‘as-a-service’ business models, e.g. car manufacturers who can adjust car specifications at consumer request. This has a very broad spectrum of possible applications. For example, digital twins help engineers determine the effects of adjustments to a particular wind turbine, pharmacists determine what the effects of a particular drug are on the human body and how it relates to other drugs and chemists develop new materials. As our digital realities match our physical one better and better, digital twins will play an increasing role developing and managing new products and processes.
