预测齿轮泵泄漏量的变化趋势有助于定量分析其性能退化过程.变分模态分解(Variational Modal Decomposition,VMD)方法对齿轮泵原始泄漏量数据进行变分模态分解,得到本征模态函数IMF,提出一种结合麻雀优化算法(Sparrow Search Algorithm,SSA)和长短期记忆神经网络(Long-Short Term Memory,LSTM)的模型,建立VMD-SSA-LSTM模型预测齿轮泵泄漏量的变化情况,并对每一个分量进行单独预测,最后将预测结果进行叠加,获得完整的预测结果.通过对比不同时间段预测结果可知,VMD-SSA-LSTM模型较单一的LSTM模型预测结果的平均相对误差最高可减小25.2％,能够完成对泄漏量的有效预测.研究结论可为齿轮泵性能衰退的定量预测提供理论支持.
Prediction of Leakage in External Gear Pump Optimized by Sparrow Search Algorithm
Predicting the change trend of gear pump leakage can help quantitatively analyze its performance degradation process.The variational modal decomposition method is used to perform variational mode decomposition on the original leakage data of the gear pump,and the intrinsic mode function IMF is obtained.A model that combines sparrow search algorithm and long-short term memory is proposed.The VMD-SSA-LSTM model is established to predict the changes in gear pump leakage,and each component is predicted separately.Finally,the prediction results are superimposed to obtain the complete prediction results.By comparing the prediction results in different time periods,it can be seen that the VMD-SSA-LSTM model can reduce the average relative error of the prediction results by up to 25.2％compared with the single LSTM model,which can effectively predict the leakage quantity.The research conclusions can provide theoretical support for quantitative prediction of gear pump performance degradation.
国家科技期刊平台
NSTL
万方数据
