LAUSR

Abstract The widespread development and application of virtual sensors in buildings will play an important role in promoting industry change and achieving smart buildings. In this study, a virtual monitoring method of pump flow rate was developed and verified. First, the pump flow parameter model was constructed based on pump operating parameters. Then, the calibration process of the model and the identification process of the pump-related efficiency were given. Finally, the experimental platform and the actual project were used to verify the virtual monitoring method of pump flow. The results showed that for the experimental test, the mean relative errors (MREs) of the three proposed methods are less than 1.79 %; for the actual project, the MREs are less than 2.48 %. Compared with existing methods, the prediction accuracies of the proposed methods have been significantly improved, and the methods are suitable for different application scenarios of pumps (with and without VFDs installed). Meanwhile, the parameters required for model calibration of proposed methods are easier to obtain. The introduction of proposed virtual monitoring methods will contribute to the replacement or backup of existing flow meters, efficiency analysis of pump systems, system fault detection and optimization control, and implementation of smart building systems.