IoT Based Fault Diagnosis of Four Stroke Petrol Engine Gearbox Using Machine Learning Approach

Abstract

It is possible to tell whether a piece of machinery is in excellent or bad mechanical newlinecondition by measuring a variety of characteristics linked to its mechanical state, including newlinevibration, bearing temperature, oil pressure, oil debris, and performance. Because of the newlinerequirement to improve machine dependability and reduce the possibility of production loss due newlineto machine breakdown, machine condition monitoring is becoming more and more important in newlineindustry. newlineThe goal of the research is to enhance the internal combustion engine assembly process newlinefault detection system for the engine manufacturing industry and to investigate the potential of newlineusing machine learning, multi-body dynamic simulation, and signal processing techniques to newlinedetect assembly faults in an IC engine. The majority of gearbox transmission system flaws are newlinecaused by wear, excessive loads, backlash, eccentric teeth, fatigue, insufficient lubrication, and newlineoccasionally manufacture After the IC engine has been assembled, a variety of faults may occur, newlineincluding crankshaft misalignment, defective bearings with problems in the cage, rolling element, newlineouter ring, and inner ring, defective gears used in the gear box with problems like broken teeth, newlinecracked or chipped teeth, and mechanical looseness. newlineThe condition monitoring of an IC engine based on a vibration signal has been researched newlinein the current work on a four-stroke engine gearbox. For the analysis, a four-stroke, singlecylinder, spark-ignition engine (BA Bajaj calibre BC-100) has been selected. The vibration signals newlineof the gear and bearing in the gearbox were recorded in both normal (baseline) and abnormal newlinesituations. The research was done in three stages. newlineUsing traditional signal processing methods like time domain, spectrum analysis, and newlineadvanced signal processing techniques like cepstrum and wavelet plots, the first phase looks into newlinedefect detection of gearboxes. The findings demonstrate that post fault detection may benefit from newlinesignal processing approaches. newlineinformation. The employment of vario