Abstract
As a liquid transportation equipment, centrifugal pumps find extensive use in industrial process control. Certain environments involving high temperature, high pressure, and highly corrosive media may not permit the installation of flow and pressure sensors. Consequently, accurately estimating the operating status of centrifugal pumps in the absence of sensors has become a prominent research focus. However, the precision of existing sensorless estimation methods remains a concern. This study introduces a HSSA-BPNN model utilizing valve opening degree, motor rotational speed and shaft power as input variables. A hybrid strategy improved sparrow search algorithm (HSSA) is developed to optimize the back propagation neural network (BPNN). Experimental results demonstrate a low mean absolute percentage error (MAPE) of the test data, which are 2.2741% and 1.6502%, respectively. The absolute errors of flow rate and head are below 0.3 m3/h and 0.5 m, respectively, with relative errors less than 4% and 3%, meeting practical applications.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
