Study Of Condition Description Method For Tribosystem Based On Tribology And Dynamics

Tribology is defined as "the science of interacting surfaces in relative motion". The tribology behaviors happened in tribosystem shows some characteristics, such as the sensitivity on the initial condition, instability and randomicity in the running process, abnormity in time and in the space. All of these features support that tribosystem is a very complicated system. Its complexity decides its diversity in its description. Therefore, it is necessary to use system-analysis method in the study of tribosystem description and in the solving problems.The purpose in this thesis is to apply the system-analysis method to tribosystem. At first, the tribology information and dynamics information were collected from the trial equipments and analysis. Then a quantitative description model of tribosystem was established. Based on the comparing of the characteristics between of tribology information and dynamical information, the description intelligence of tribosystem was available in some degree.The main content in this thesis is as following:1) In the trial, the tribology and dynamical information were collected in synchronous and processed from pin-on-disk tester (MMW-1) and rolling bearing tester (RBT). Based on the analysis of two tribosystem, the tribology simulation model and dynamic model were proposed.2) From the connotion of tribosystem, it is pointed out that the study of tribosystem should be deployed through its system features and its condition features. The condition features is decided by the system features. To obtain exact condition discerning result, the condition features should be examined with system feature. From information angle the correlation between tribology and dynamic information was discussed.3) Because of the complexity, the incompatibility and the contradiction in tribosystem, the extenics principle was applied to describe tribosystem condition. Based on the establishment of the evaluation matter of four tribosystem wear conditions (including such as normal wear condition matter etc.), the relationshipvalues between unknown condition and each of four wear conditions were calculated respectively. The biggest value in four results means the unknown condition is closer to this wear condition than to others. This evaluation method changed the traditional qualitative description to the quantitative and multi-feature description. Using same principle, the classification quantitative research was used on fault type's diagnosis in RBT. The results show that this method is practical.4) Case based reasoning can utilize the practical cases to help the diagnosis when some rules are difficult to obtain or the rules have strong subjective. Cases should be the main part in tribosystem knowledge base. In case base from the trial data, the fuzzy similar priority relationships between the object case and original case were established through comparing the factors in case. Then the most similar original case was found out with the object case after the similar order was obtained.5) With rough set theory, the uncertainty of tribosystem information was processed. Under the case base reasoning, the increase of case will induce the contradiction among cases and incompatibility. In this thesis the knowledge in case base was classed and reduced with rough set principle to decrease redundancy data, enhance the reasoning exact.