【摘要】 变速变排量液压驱动单元是一种泵控液压驱动单元，可以通过改变电机转速和泵排量调整其输出流量，提高输出流量和液压执行器需求流量的匹配程度，同时提高电机和液压泵的能效。文章针对变速变排量液压驱动单元能效模型存在难求解和不精确的问题导致相应节能控制策略难以实现的情况，基于机器学习算法提出一种能耗建模方法，实现驱动单元的能耗预测；基于实验平台采集的能耗数据，采用6种独立的机器学习模型(岭回归、支持向量机(support vector machine, SVM)、随机森林、梯度提升决策树(gradient boosting decision tree, GBDT)、极端梯度提升(extreme gradient boosting, XGBoost)、逆向传播神经网络(back-propagation neural network, BPNN))和1种堆叠模型进行能耗建模，并采用均方根误差(root mean squared error, RMSE)、平均绝对百分比误差(mean absolute percentage error, MAPE)和决定系数R~2对模型进行评估。结果表明，使用集成学习中堆叠模型构建的能耗预测模型在测试集中有最好的预测精度。更多还原

【Abstract】 The variable-speed variable-displacement pump unit is a pump-controlled hydraulic drive unit, which can adjust the output flow by changing the motor speed and pump displacement, thus improving the matching degree between the output flow of the pump and the required flow of the actuators, and promoting the energy efficiency of the motor and hydraulic pump. However, the energy-saving strategy of the variable-speed variable-displacement pump unit is difficult to realize because of its inaccurate energy model. Based on the machine learning algorithm, an energy consumption modeling method which can predict the energy consumption of the drive unit is proposed. The energy consumption models are respectively built by six types of independent machine learning models, namely ridge regression, support vector machine(SVM), random forest, gradient boosting decision tree(GBDT), extreme gradient boosting(XGBoost) and back-propagation neural network(BPNN), and a stacked model based on the energy consumption data collected by the experimental platform. The models are evaluated with root mean squared error(RMSE), mean absolute percentage error(MAPE), and decision coefficient R~2. The results show that the energy consumption prediction model based on the stacked model algorithm has the best prediction accuracy in test set.更多还原