随着柴油机高喷油压力和燃油后喷技术的应用，润滑油受燃油稀释的问题愈发严重，同时温度对燃油黏度有显著影响，建立考虑温度和燃油稀释的黏度模型，对润滑油性能研究和柴油机摩擦副模拟计算具有重要意义。以柴油机用美孚15W-40润滑油为例，通过试验分别测定了未混入燃油的原15W-40润滑油和混入燃油体积分数为3%，6%，9%，12%，15%的15W-40润滑油在20~180 ℃之间的黏度变化。构建了考虑燃油体积分数的Vogel方程、Andrade方程及Reynolds方程，并比较了其拟合效果。构建了量纲一的稀释黏度比温度方程，反映了燃油稀释在不同温度下对黏度的影响程度和规律。构建了基于润滑油温度和混合燃料体积分数的反向传播神经网络黏度模型。结果表明：Vogel方程对15W-40润滑油被稀释前后黏温曲线拟合效果最好；所构建的反向传播神经网络黏度模型对15W-40润滑油黏度预测值中90%的误差小于2%，整体误差不超过6%。
With the application of high-pressure fuel injection and fuel post-injection technology of diesel engine, the dilution problem of lubricating oil caused by fuel becomes more and more serious, and temperature has a significant effect on the fuel viscosity. Therefore, establishing a viscosity model considering temperature and fuel dilution is of great significance to the study of lubricating oil performance and the simulation of diesel engine friction pairs. Taking Mobil 15W-40 lubricant of diesel engine as an example, the viscosity changes of different lubricants between 20 ℃ and 180 ℃ were measured by experiment including 15W-40 lubricant and its blends with 3%, 6%, 9%, 12% and 15% volume fraction of fuel. Vogel, Andrade and Reynolds equations considering fuel volume fraction were built and their fitting effects were compared. The dimensionless dilution viscosity ratio-temperature equation was constructed to reflect the influence degree and law of fuel dilution on viscosity at different temperatures. A back-propagation neural network viscosity model based on lubricating oil temperature and mixed fuel volume fraction was finally constructed. The results show that Vogel equation has the best fitting effect for the viscositytemperature curve of 15W-40 lubricating oil before and after dilution. For the predicted viscosity of 15W-40 lubricating oil of back-propagation neural network viscosity model, the error is less than 2% among the range of 90% and does not exceed 6% in a whole.
