Please use this identifier to cite or link to this item:
|Title:||FUZZY LOGIC MODELING OF WEAR IN FERROUS ALLOYS|
|Authors:||Krishna, Chavan Gurling|
FUZZY LOGIC MODELING
METALLURGICAL AND MATERIALS ENGINEERING
|Abstract:||Wear is one of the main ways, by which inanimate objects loose there usefulness. Wear is inevitable, when two surfaces slide against each other. In the present study, dry sliding wear against the counterface of hardened steel has been modeled through uncertainty based fuzzy logic modeling for chromium steel and plain carbon steel. Initially wear data pertaining to the relation between input and output variables have been collected from open literature and analyzed to formulate template based model. The selected input variables are normalized pressure, sliding velocity, chromium content, relative humidity and carbon content. The wear has been characterized in terms of non dimensional wear coefficient. The data collected do not belong to a single unified base reflecting variation of all the selected input variables and so, two fuzzy models have been constructed. In one model the effect of normalized pressure, sliding velocity and chromium content has been modeled at constant relative humidity to study wear in chromium steel. The effect of relative humidity and carbon content has been modeled in a second model, for plain carbon steel sliding under a fixed normalized pressure and sliding velocity. MATLAB software has been used for fuzzy modeling which yield the plot selected input variables and the output variable of wear coefficient. Finally, the output results and observed experimental results are compared to check the validity of each model. Statistical analysis of correlation between the experimental data and model output has been used to test validity of the models. The first fuzzy wear model for chromium steel at constant relative humidity shows deviation as indicated by correlation coefficient of 0.6. The fuzzy wear model for plain carbon steel including the effect of relative humidity on wear gives a better matching with correlation coefficient of 0.93. It is interesting to note that the first model for chromium steel has indicated transition in wear from mild to severe, induced by load and sliding velocity. The model for plain carbon steel has shown that at higher relative humidity there is transition to mild wear and the level of relative humidity which causes this transition and depends on the carbon content of plain carbon steel. Thus, it is evident that it is very important to indicate the relative humidity of the environment while performing experiments in wear|
|Appears in Collections:||MASTERS' DISSERTATIONS (MMD)|
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.