Study On The Identification Of Lubricant Additives And Acid Value Determination Method Based On Mid Infrared Spectroscopy

As we all know,mechanical equipment is bound to be accompanied by friction and wear on the contact surface of relative motion.If there is no targeted method to reduce the friction,then the life of the machine will be very short.With the development of The Times,lubricating oil technology is more and more complete,coupled with the wide application of additives,lubricating oil has become an important product.The basic principle is simply to block the two friction surfaces.The lubricant can be firmly adsorbed on the friction pair of the machine parts and form a layer of oil film.The oil film and the friction surface of the machine parts are strongly engaged,and the two friction surfaces are therefore separated.This principle successfully converts the friction of the equipment into the friction between the molecules of the lubricating oil itself,thus reducing friction and wear.Equipment lubrication is a key part of equipment maintenance.Either oil deterioration or oil reduction will increase equipment wear.In this context,the importance of equipment lubrication is self-evident.Lubricating oil is composed of base oil and additive.Base oil is the main part of lubricating oil,which determines the basic properties of lubricating oil.The additive is the essence of lubricating oil.When it is compatible with base oil,it can improve and enhance its physical and chemical properties.Therefore,additives are rich in information to characterize the main performance of lubricating oil,and it is also the key to ensure the quality of lubricating oil.Different requirements of equipment lubricants,additives are also different.The type of additive determines the classification of the oil,and also determines the performance of the oil.How to test its performance scientifically is of great significance.Practice has proved that the physical and chemical indexes also define the properties and health status of lubricating oil.Oil monitoring technology of equipment lubricating oil in use can predict oil pollution and abnormal wear in advance by regularly analyzing the composition and performance index changes of lubricating oil in use.It can not only timely grasp the use status of lubricating oil,confirm whether the lubricating oil meets the requirements and can be used normally,but also act as an intermediary to indirectly obtain the wear of mechanical equipment and discover problems in advance Solve the problem in time,reduce the mechanical failure,reduce the maintenance cost,improve the service life of the equipment.Therefore,the qualitative analysis and physical and chemical indexes test of equipment lubricating oil additives have very important theoretical and practical significance.Aiming at the above problems,this paper studies two parts respectively from the perspective of rapid test of lubricating oil performance1.In this paper,several classical classification algorithms,including support vector machine(SVM),are studied to identify the small amount of additives in lubricating oil of near category mechanical equipment In order to optimize the characteristic bands of mid infrared spectrum,a multi-class classification model was established by embedding genetic algorithm(GA)and sparrow search algorithm into(OVR SVMs),decision tree(DT),gradient lifting decision tree(GBDT),random forest(RF).2.For the rapid test method of physical and chemical indicators,considering that the acid value can characterize the aging quality of lubricating oil,and when the acid value exceeds 2,the oil can not be used any more,so this paper takes the wind turbine gear oil as an example to find a nondestructive method to quickly determine the acid value of the wind turbine gear oil.In addition,some classical regression algorithms,decision tree,support vector regression machine and other feature selection methods are also used to establish the regression prediction model between the physical and chemical performance index data and acid value quality index of wind turbine gear oil,so as to directly use the infrared spectrum data of oil for acid value determination.The test results show that the recognition rate of the new model with OVR SVMS and GBDT embedded GA-SSA optimization technology can reach 100%,and eight kinds of lubricating oil with different additive contents can be identified simultaneously.In the method study of acid value index determination,the model constructed by combining the sparse representation of L1 with the decision tree DT performs well,and the determination coefficient reaches 1.The above results prove the feasibility of the method studied in this paper.