The Concept of Virtual Sensors in a Digital Twin Framework

Sammanfattning

As digitalization is becoming a hot topic within the construction industry, the sector faces some barriers in adopting new technologies, e.g., by being fragmented between all involved stakeholders, low knowledge transfer between projects etc.. With digitalization, new opportunities arise to mitigate such barriers. An example is the possible adoption of digital twins into the sector. This relatively new digitalization technology represents a promising concept, in academic research literature it is described as a useful possibility in the construction sector's road to industry 4.0, especially as it is embodying a convergence of various technologies to facilitate enhanced project management and increased efficiency. It can greatly mitigate the issue of a fragmented industry with increased communicative opportunities. To get an effective use of this concept it is imperative to have a solid foundation in data and components to derive all valuable information from projects. As it encapsulates multifaceted categories, such as Artificial Intelligence (AI) and Machine learning (ML), the concept of digital twins can often be viewed as costly and advanced. This is often derived from the lack of data available to effectively implement these technologies. Though in recent years the technical advancements and developments of Information, and communication technologies (ICT), and Internet-of-things (IoT) sensors, has brought forth a rise in data monitoring for smart buildings and the indoor environment. Such advancements can often come with extraneous costs, unrelated to an already high cost industry. Having sensors connected to a building is a future step in total control and monitoring of our future structures, but realizing alternatives such as, in the case of sensors, virtual sensors (VS) may increase the sustainability of our future construction industry. A critical inquiry therefore arises regarding the feasibility of implementing alternatives to so-called traditional systems. This work evaluates the concept of VS in regards to possibly alleviating some issues with data collection from traditional systems and sensor assets. A connection is made from digital twins in construction to the development of VS. The general idea is that an integration of VS into digital twins can enhance monitoring, simulation, and performance forecasting. However, accuracy compared to physical sensors (PhS) may vary based on the specific use case. The study shows that with simple means it is possible to replace some PhS with VS, but that the need for PhS is still prominent. By instead focusing on creating cyber-physical systems, where VS are nodes in the network. VS and PhS can increase the validity and reliability of the entire system, when co-creating data generation.