Future Landscape of Digital Twins
- In general, three components are required in the development of digital twins: the software, hardware, and middle ware components.
- The hardware components required for the development of digital twins include sensors and actuators, while the software components include data dashboards and modeling and simulation software packages.
- One gap/area of concern for business owners in the use and implementation of digital twin technology is model accuracy.
This research provides an overview of the technology infrastructure (software and hardware) required to develop digital twins and a brief description of each infrastructure component. Also, some gaps/areas of concern for digital twins and the associated technologies are provided. However, information on such gaps/areas of concern for general digital twin systems is limited in the public domain. Hence, the research team pivoted to provide industry/use-case specific gaps in using digital twins and the associated technologies. Details of the research strategy employed in this research are given in the research strategy section below.
Digital Twin Technology Requirements
- The components that make up a digital system may vary by application. However, three general components make up a digital twin infrastructure: hardware components, software components, and middle ware components.
Digital Twin Hardware Requirements
- The hardware infrastructure requirements for digital twins include Internet of Things (IoT) sensors, actuators, routers, and edge servers.
- Sensors are devices that collect physical parameters/characteristics such as temperature, speed, or humidity and turn them into digital signals. The sensors then send these digital signals to IoT devices. Examples of sensors are meters, thermometers, and probes.
- Actuators are devices that translate the digital signals from an IoT system into different physical equivalent actions. For example, an actuator can translate a digital command from a controller into motion in a robotic arm.
- Routers transmit network signals from one device or location to another. Edge servers control local data management and manipulation at a device level.
Digital Twin Software Requirements
- The software infrastructure for digital twins includes analytics engines, machine learning models, data dashboards, and modeling and simulation software for design.
- Analytics engines leverage raw observations and records to generate valuable and actionable business insights. Analytics engines often include machine learning models that learn from previous observations to represent and predict the future behavior of the system or its components.
- Data dashboards display a live view of observations from different components of the system. They enable real-time monitoring of the system.
- Modeling and simulation software are tools used to replicate the physical parameters and behaviors of a physical system in a digital form.
Digital Twin Middleware Requirements
- The middle ware acts as a bridge between software and hardware. At a fundamental level, it comprises a central data repository that holds data from different sources. The tasks handled by middle ware components may include connectivity, data processing, data integration, data visualization, data quality control, and data modeling and governance. Most of these tasks are usually integrated into IoT platforms. Hence, IoT platforms form a crucial part of the middle ware.
Digital Twin Technology Gap/Area of Concern #1: Model Accuracy
- According to Challenge Advisory, the biggest concern held by many business owners who want to implement digital twin technology is “the risk of misrepresenting the object or system they want to replicate using this technology.”
- Similarly, business owners are concerned about the accuracy of the simulation that will result from the digital twin. That is, how well the digital twin will represent and predict the outcome of the physical system under different conditions.
- According to consultants at Challenge Advisory, the concern of model accuracy is addressed by using specific sensors that monitor a physical object to create a digital equivalent, and even a representation of the inner workings of the physical object. To generate a representation of the inner workings of an object, physical splitting may be required. The split object is then represented manually using animations and computer code.
- In many cases, such as in the representation of processes and systems, dedicated digital twin software is required for building the required outcomes. In such cases, a cycle of building and manual revisions are required to ensure the accuracy of the final product.
DT Technology Gap/Area of Concern #2: Sensors and Instrumentation, and Modeling in Nuclear Power Plants
- Some challenges that create gaps in the application of digital twin technology to nuclear power plant operation and design are problems with real-time data collection and integration, and the challenges associated with the operating environment.
- In nuclear power plants, advanced sensors and instrumentation infrastructure are limited by problems of slow speeds in data acquisition, integration of data streams from different sources, and integration with digital twin technologies.
DT Technology Gap/Area of Concern #3: Outdated Systems of Data Capture in Assembly Lines
- Factories and assembly lines that still rely on old methods for capturing time and motion data, such as pen, paper, and stopwatch approaches for data capture, may find it difficult to incorporate digital processes and hence leverage the digital twin technology.
- For such conditions, more modern systems of real-time data sharing and scaling are required. These modern methods can also provide visual representations of the state of assembly line processes. An example of such a system is Drishti.
To answer this request, the research team consulted industry publications such as those from AltexSoft, a consulting company that builds custom software solutions, and Challenge Advisory. The research team could provide the technology (software and hardware) requirements for building digital twins. However, there was limited information on the general gaps/areas of improvement for digital twins and the associated technologies. This is because of the limited adoption of digital twin technologies in commercial processes.
To address the above problem, the research team pivoted to identify gaps and areas of concern in using digital twin technology for specific industries. Through this approach, the research team could identify some gaps in using digital twin technology for assembly lines and nuclear power plants.