Prediction of Energy Consumption Characteristics of Electro-Proportional Piston Pumps Based on GA-SVR

To optimize and control hydraulic pumps in an energy-efficient manner, it is crucial to predict their energy consumption accurately. This paper proposes a method for predicting the energy consumption characteristics of an electric proportional piston pump using support vector regression (SVR) and genetic algorithm optimization. The method uses test signals to extract the characteristic variables that affect energy consumption and develops a machine learning model based on SVR and input features to predict the energy consumption parameters of the pump. The experimental results demonstrate that the proposed method achieves excellent performance on the test set of electric proportional pump energy consumption data. The R2 coefficient of the model reaches above 0.99 for both pressure and flow prediction, and the RMSE and MAPE are also low, indicating that the developed model has good generalization ability and high prediction accuracy for the energy consumption characteristics of electric proportional piston pumps.