Detection of Oil and Filter Feasibility in Ship Gearbox with SVM (Support Vector Machine) Method Based on Microcontroller

Abstract

Lubrication systems on ships are vital for a ship to operate while sailing. Lubrication serves to prevent direct friction between components on the ship engine, especially the gearbox. Gearbox is one of the components of the ship crane that acts as the propeller of the ship. Obviously related to friction between components, therefore a lubrication system is needed. Some lubrication systems on Indonesian warships still use manual systems to monitor oil and filter change schedules. This can make the operator forget about the oil and filter change schedule. With the development of technology that is now the author makes a prototype with an oil and filter feasibility detection system on a ship gearbox. The author uses the SVM method that is able to classify properly or not like an oil. By using a viscosity sensor and a temperature sensor to be processed by the SVM method and a color sensor to detect a color in the oil filter. The results of the SVM method are quite accurate with 100% accuracy can distinguish good and bad oil from the samples used are new oil and used oil. There is a change in the intensity of the light that causes the color sensor readings in conditions of 100%, conditions of 30%, conditions of 40%, and conditions of 50%.