A Smart Metrology Laboratory Environmental Monitoring and Recognition Method Based on Digital Twin and Improved Transformer

ABSTRACT Aiming at the problems of poor reliability and low visualization in the monitoring of power metrology laboratories, a smart power metrology laboratory environmental monitoring and recognition method based on digital twins and improved Transformer is proposed. Firstly, a smart power metrology laboratory architecture is designed based on digital twin technology, achieving comprehensive perception of the laboratory through the interaction between physical entities and virtual platforms. Then, the large kernel block module is added to the Swin Transformer model to enhance the representation ability between cross—modal feature information, and the improved Swin Transformer model is deployed in the intelligent layer of the laboratory architecture. Finally, the improved model is used to analyse the laboratory monitoring information, and multi‐source information is fused for decision‐making to obtain reliable identification results of personnel behaviour and metrology device status. Experimental analysis of this method based on a virtual smart power metrology laboratory scenario shows that this method can effectively identify various behaviours of personnel. The recognition accuracy of sleeping behaviour reaches 96.74%, and the average recognition accuracy of metrology device status is 91.74%. Meanwhile, the study also points out the limitations of the current system in detecting some failure modes, providing more comprehensive theoretical and technical references for the construction of smart power metrology laboratories.