Research On Compatibility And Performance Prediction Of Lubricating Oil Additives

Lubricating oil is composed of base oil and additives.It is used in various mechanical equipment to play the role of lubrication,anti-wear,antifriction,cooling,rust prevention and shock absorption.The function of oil all based on the basic performance of base oil and the function of additives.The base oil decide the basic function of the oil,such as the condensation point,boiling point,density,etc.Additives can give the base oil some properties that it did not have or strengthen some of its properties.Synergism and mutual resistance may occur between different additives,so the selection and compatibility of base oil and additives has always been a difficult point in lubricant formulation design.In this paper,GA-BP neural network is used to fit the existing experimental data,so as to quickly and accurately calculate the lowest function of oil and additives.My research content is as follows:(1)I50SN mineral oil was used as base oil,molybdenum oxydialkyldithiop hosphate(MoDTC),zinc dialkyldithiophosphate(T202),acid aliphatic phosphate(T308),trimethyl phosphate(T306)and nitrogen-containing derivative of sulfur and phosphorus(T305)were used as additives to investigate the antiwear and friction-reducing properties of oil in one factor conditions,and the element di stribution of the protective film on the wear spot surface was observed and an alyzed,and its antiwear and antifriction mechanism was deduced.(2)Five additives,MoDTC,T202,T308,T306 and T305,were added to 150SN mineral base oil,and their compatibility was studied.A series of experimental results show that MoDTC(0.3%)+T202(1.2%)+T308(0.2%)+T306(0.4%)has the best lubrication performance.The energy spectrum analysis of the wear spot surface infers that the surface is mainly formed with protective film containing FePO4,MoS2,etc.,which plays an anti-wear and antifriction role.(3)Based on algorithms such as artificial neural network,GA-BP neural network and genetic algorithm,use five additives in different samples as the answer,the friction coefficient and wear scar in the friction test as the label,taking the proportion of five additives in the composite additive formula as the input,and the index of the friction coefficient and wear scar width as the output,the model of machine learning is built to predict the best formula of lubricating oil additives,And obtained experimental verification.