Extension of the system boundary of the Digital Twin onto the sensors of the Physical Twin through the introduction of redundant soft sensors

Abstract

The benefit of Digital Twins depends to a large extent on the quality of the sensor data provided. In many cases, sensor failures are only detected late in operation which can lead to serious consequences. For this reason, one approach to reduce the resulting safety issues is to use redundant sensor systems that monitor the same measureand. However, due to the additional sensors required, this is associated with additional financial and design effort.In this publication an alternative strategy is presented, which provides a redundant sensor system with the help of soft sensors. Soft sensors use already installed physical sensors to anticipate a new measured variable via algorithms. They are often used to avoid placing sensors in inaccessible locations, but are used here to perform redundant computation of already existing metrics. The sensor data of physical and soft sensors are used as input variables for a Digital Twin. Here, these are compared with each other and can be critically questioned by the twin itself. This makes it possible to extend the system boundary of the Digital Twin to the sensors themselves and provided input variables can be checked for their validity. This allows sensor failures to be detected at an early stage and consequential damage to be averted.