摘 要

本文为车用润滑油的质量检测提出了一种更为简便、高效、低成本，测量结果更为准确的新方法，利用光谱反射法来对润滑油进行质量检测。对比国外润滑油的检测和更换方法，在国内，润滑油的更换并不能准确的根据本身的质量、工作效能来进行，这样不仅加重发动机内部件的磨损，影响发动机寿命，同样也造成资源的浪费。在本文中通过对润滑油的种类和衰变因素的研究，得出润滑油的评价指标，设计并进行实验。同时对相关光学知识进行学习和研究后，利用MATLAB软件，建立反射光谱的仿真模型，构建目标函数，结合遗传算法对实验数据进行优化拟合，得到不同润滑油样品的反射率、消光系数等与波长的变化关系，探究随着发动机磨损量的增大，润滑油发射率的变化范围和趋势。通过实验以及程序拟合的结果分析，该方法能较好的反映润滑油质量的变化，一种可行性高的检测方法。

关键词：汽车；发动机；润滑油；光谱反射法；色散模型；优化算法

Abstract

In this paper, a spectral reflection method is proposed to detect the quality of vehicle lubricating oil，which is simpler, more efficient, lower cost and more accurate. Comparing with the inspection and replacement methods of lubricating oil abroad, the replacement of lubricating oil can not be carried out accurately according to their own quality and working efficiency at home, which not only aggravates the wear of internal parts of the engine, affects the life of the engine, but also causes the waste of resources. In this paper, through the study of the types and decay factors of lubricating oil, the evaluation index of lubricating oil is obtained, and the design and experiment are carried out. At the same time, after studying and researching the relevant optical knowledge, the simulation model of reflective spectrum is established by using MATLAB software, and the objective function is constructed. The experimental data are optimized and fitted by genetic algorithm. The relationship between reflectivity and wavelength of different lubricating oil samples is obtained. With the increase of engine wear, the range and trend of lube oil emissivity are explored. Through the analysis of the results of experiments and program fitting, this method can better reflect the quality change of lubricating oil, and it is a feasible detection method.

Key words: Automotive Engine；Lubricating Oil；Spectral Reflectometry；Dispersion Model；Optimization Algorithm

Contents

摘要 III

Abstract IV

1 Introduction 1

1.1 Background 1

1.2 Status at Home and Abroad 2

1.3 Main Research Contents and Significance 5

2 Engine Lubricating Oil 6

2.1 Engine Operating Characteristics 6

2.2 Lubricating Oil Effects 7

2.3 Performance Requirements for Engine Oils 10

2.4 Types and Selection of Lubricants 13

2.4.1 Types of Lubricants 13

2.4.2 Lubricating Oil Selection 14

2.5 Engine Lubricating Oil Decay 16

2.6 Lubricating Oil Replacement Standard 18

2.7 Lubricating Oil Performance Evaluation Method 20

3 Spectral Reflection Method 23

3.1 Single Layer Film Reflectivity 23

3.2 Dispersion Model 27

3.2.1 Dispersion of Light 27

3.2.2 Cauchy Dispersion Model 27

3.2.3F-B Model 28

3.2.4Tauc-Lorentz Model 29

4 Optimization Algorithm 30

4.1 Optimization Principle 30

4.2 Simulated Annealing Algorithm 30

4.2.1 Source of Simulated Annealing Algorithm 30

4.2.2 Optimization Process of Simulated Annealing Algorithm 31

4.2.3 Metropolis Acceptance Criteria 31

4.2.4 Termination Criteria for Simulated Annealing Algorithm 32

4.2.5 Characteristics of Simulated Annealing Algorithm 32

4.3 Genetic Algorithm 33

4.3.1 Source of Genetic Algorithm 33

4.3.2 Some Concepts in the Algorithm 34

4.3.3 Stop Criterion of Genetic Algorithm 35

4.3.4 Optimization Process and Algorithm Idea of ​​Genetic Algorithm 35

4.3.5 Characteristics of genetic algorithms 37

4.3.6 Large Variation and Adaptation of Genetic Algorithm 38

4.3.7 Operational Points of the Genetic Algorithm 40

4.4 Application of Genetic Algorithms in MATLAB Software 41

4.4.1 Introduction of MATLAB 41

4.4.2 Genetic Algorithms in MATLAB Software 42

5 Experimental Spectrum Simulation Fitting and Optimization Program 44

5.1 Spectral Modeling and Simulation 44

5.2 Objective Function Establishment 46

5.3 Optimization of the Objective Function 47

6 Experimental Research 49

6.1 Experimental Protocol 49

6.1.1 Instruments and Reagents 49

6.1.2 Preparation before the Experiment 51

6.2 Experimental Data Collection and Processing 53

6.2.1 Experimental Data Collection 53

6.2.2 Data Fitting Processing 54

6.3 Conclusion 59

7 Summary and Outlook 61

Reference 64

Appendix 68

Appendix A MATLAB Program for Reflectivity Function 68

Appendix B MATLAB Program for Object Function 68

Appendix C MATLAB Program for Optimizing Fitting 69

Appendix D MATLAB Program for Inversion of Reflectivity 70

Acknowledgements 73

1 Introduction

1.1 Background

The lubricating oil forms an oil film on the friction surface of the inner wall of the engine cylinder to achieve liquid friction, which can reduce frictional resistance, reduce power consumption, and reducing component wear. Secondly, Lubricating oil can also clean metal debris from friction surfaces and deposit carbon on cylinders, pistons, piston rings, etc. It can also play the role of sealing and leak-proof ^[1].

According to the typical automobile energy distribution measured by the Environmental Protection Agency, The engine friction loss accounts for 4.5% of the total energy loss caused by friction, while the failure caused by friction accounts for 47.2% of the total engine failure ^[2]. It can be seen that the lubricating oil plays an important role in the performance of the engine. Therefore, by obtaining the lubricating oil film data, the working condition and working life of the engine can be known in some respects.

In the research on the detection of engine lubricating oil, the conventional testing method uses different monitoring devices and instruments to measure the performance change of lubricating oil and the information of wear particles, and analyze and judge the quality of lubricating oil ^{[3] - [5]}. Therefore, based on saving energy, the purpose of replacing the lubricating oil in time, reducing the friction fault of the engine and prolonging the service life of the engine. However, the above methods all have inconvenient measurement, high measurement cost, and accuracy is also affected by many factors.