基于LS-SVM的柴油机润滑油中磨粒含量预测

Forecasting of Wear Particle Concentration in Diesel Engine Lubricating Oil by Least Squares Support Vector Machine

介绍了最小二乘支持向量机(LS-SVM)回归算法的基本原理,并以490BPG型柴油机润滑油中磨损磨粒为研究对象,使用LS-SVM对磨粒的浓度数据进行了回归拟合并预测,并与基于人工神经网络的预测模型的预测结果进行了比较。结果表明,LS-SVM的预测模型的精确度较高,泛化能力强,是用于润滑油中磨粒浓度预测的一种有效的方法。

The basal principle of least squares support vector machine （LS-SVM） regression algorithm was introduced, and the wear particle concentration in 490BPG diesel engine lubricating oil was fitted by regression analysis and the wear tendency was predicted by LS-SVM. The predicted result was compared with that by artificial neural network （ANN） model. The result shows that the LS-SVM has better integrative performance and generalization ability and high precision, so it is an effective method for being used in forecasting of wear particle in lubricating oil.