Mengesan tahap kelikatan minyak pelincir dalam kenderaan menggunakan sistem logik kabur

Maintaining the quality of lubricant oil quality can guarantee maximum ability in engine functions of vehicles. Currently, the quality of lubricant oil is primarily determined by two factors, namely, vehicle’s mileage and duration. However, these judgments are inaccurate because there are many oth...

Full description

Bibliographic Details
Main Authors: Norsalina Harun, Siti Norul Huda Sheikh Abdullah, Khairuddin Omar, Siti Rozaimah Sheikh Abdullah
Format: Article
Language:English
Published: 2006
Online Access:http://journalarticle.ukm.my/1453/
http://journalarticle.ukm.my/1453/
http://journalarticle.ukm.my/1453/1/2006-12.pdf
Description
Summary:Maintaining the quality of lubricant oil quality can guarantee maximum ability in engine functions of vehicles. Currently, the quality of lubricant oil is primarily determined by two factors, namely, vehicle’s mileage and duration. However, these judgments are inaccurate because there are many other factors like conductivity, humidity, temperature and viscosity that may affect the oil quality. In addition, improper treatment of used lubricant oil will greatly pollute the environment. From the investigation carried out, some parameters were suitably identified to determine the current quality of lubricant oil. Those parameters were error and change of error of lubricant oil temperature that were used as the inputs to a fuzzy logic system. The expert knowledge was compiled to justify the human expertise. This developed fuzzy logic system was able to function on its own by using Prolog programming language. The language eased the representation of rule-based knowledge so that its inference can be performed naturally. The obtained data of temperature relation to the lubricant oil quality were applied to the developed membership function of the the fuzzy logic system and had gone through several stages to obtain crisp values representing the lubricant oil quality. The results obtained shows that 90% of the data can be predicted with 82.4 to 98.11% accuracy