The aim of this study is to determine the condition (fresh or used) of the engine oil used in motor vehicles via an electronic nose. 50 cc fresh elf 10/40 engine oil and 50 cc 10000 km used same engine oil were sniffed by the electronic nose. A sensor block which was formed by 11 electrochemical gas sensors was used in the electronic nose prepared for this study. Sensor data were converted to digital data by using analog digital converter of the two arduino uno microcontroller development cards. Then, the digital data were transferred to the computer. The control program for recording sensors' data was prepared in LabVIEW software. After ventilation process, by putting the oil sample into the sniffing box, the odor of oil was sniffed and recorded for 120 seconds via the sensor block which is placed inside to the sniffing box. Then, various features of these data were extracted and these extracted features were classified with kNN classification algorithm prepared in MATLAB program. As a result of the performed classification, the fresh engine oil could be distinguished from the oil that was used for 10000 km in a gasoline engine from their smell with an accuracy of 95.14% by the electronic classification; gas sensor; kNN algorithm; oil quality; signal processing
Anahtar Kelimeler: Classification, Gas Sensor, Oil Quality
|
