DTs are an extension of the focus on data-driven approaches across industries and use cases. A DT is a virtual representation of a physical object known as a physical twin. Using real-time data, DTs are capable of instantly reflecting any changes in the physical twin in their digital copy. DTs differ from standard simulations–which use static data and study a single process–because they can run simulations on multiple processes. There are broadly four categories of DTs, each increasing in the magnitude of the type of asset or process being virtually represented. This includes the simplest form, known as a component or part twins, which could be as simple as extracting data from a sensor in the nuts and bolts of a particular piece of equipment, to the broadest form, known as a process twin to represent production factories virtually.

In order to build DTs, the physical twin is fitted with IoT devices to collect data (e.g., indicators of asset health and temperature), which is processed and applied to its digital copy. The data is sourced from drones, sensors, and satellite images; the DT’s accuracy to simulate the physical twin is greatly influenced by the amount of data used. Using a combination of ML models and data visualization, the DT can run simulations to understand the performance and efficiency of the physical twin and also how it changes under several real-world scenarios, including how it is performing in the present and predicting how it will perform in the future. DTs are similar in appearance to their physical counterparts (e.g., a wind turbine, a car, an off-shore oil rig, or factory equipment), and, in some instances, such as those used in product development and design, are developed before the physical object that they replicate.