Flow rate prediction of gear pump based on VAR model with lag factor

As a key power component in hydraulic systems, the gear pump is widely employed in industrial equipment due to its strong self-priming capability and high contamination resistance. However, its external characteristics are highly sensitive to variations in operational conditions owing to the viscosity constraints of hydraulic oil. Based on an in-depth analysis of flow dynamics, this study proposes a Vector Autoregression (VAR) model incorporating lag operators to predict the flow rate of gear pumps. For hydraulic systems without temperature control units, the proposed VAR model uses load pressure and ambient temperature as input variables to forecast gear pump flow rates. Through outdoor experiments on gear pumps in a railway car retarder system, the prediction results demonstrate that the proposed VAR model achieves significant improvements in prediction accuracy compared with general VAR, ARIMA, and LSTM models.