Virtual sensor architecture for indoor air quality monitoring

Human exposure to poor air quality is a leading risk factor in the Global Burden of Disease (GBD) study, estimating 22,000 premature deaths related to indoor air pollution in 2019 in Europe. Diverse pollutants are found in manufacturing environments resulting from both combustion and non-combustion sources, including Particulate Matter and Volatile Organic Compound. Internet of Things (IoT) air quality monitoring can enhance awareness and support informed decision making towards better air quality. However, hardware sensors are not always capable of monitoring particular characteristics and behaviour of a pollutant, for instance, spatial limitations may impede deploying sensors close enough to the source of the pollutant. Virtual Sensors can extend hardware sensing options via signal processing and data integration. This paper presents an architecture for training and deploying virtual sensors. A virtual sensor is implemented using the architecture in the context of additive manufacturing to estimate the production of Volatile Organic Compounds (VOCs) of 3D printers and their transfer into the rest of the space. In the case study, the 3D printers are installed inside cabinets to limit the transfer of pollutants to the exterior. Several of these virtual sensors are deployed to monitor the VOCs produced by the 3D printers and the transfer rate out of the cabinets. The paper includes some early results and initial insights on the accuracy and usefulness of virtual sensors. Virtual Sensors can be cost-effective solutions when monitoring systems are escalated by reducing number of hardware sensors and complexity.