MACHINE LEARNING-BASED FAULT DIAGNOSIS IN ENGINES USING IOT BASED THERMODYNAMIC FEATURE VECTORS

Abstract

By using real-time data from IoT sensors to
improve predictive maintenance methods, the
combination of Machine Learning (ML) and
the Internet of Things (IoT) has completely
changed the way engine problem diagnostics
is done. Consequently, this study investigates
the creation and deployment of a problem
diagnostic system based on ML that makes
use of thermodynamic feature vectors
retrieved from Internet of Things (IoT)
sensors. The combination of temperature,
pressure, and flow rates in these feature
vectors gives a complete picture of the
engine's state and allows for very accurate
defect identification and prediction. Modern
engines produce massive amounts of
complicated data, which may be
overwhelming for older diagnostic systems
that depend on rule-based procedures, human
inspections, and Diagnostic Trouble Codes
(DTC). Not only are these approaches
laborious and error-prone, but they also can't
do much in the way of preventive
maintenance. In order to overcome these
constraints, this study analyses and interprets
the high-dimensional data collected by
Internet of Things (IoT) sensors using
sophisticated ML methods such neural
networks, decision trees, and support vector
machines. Improving engine efficiency and
safety while cutting maintenance costs and
downtime is the goal of the suggested system's
real-time monitoring and predictive
maintenance. The ML models can catch little
irregularities and foresee possible problems
before they become big problems since they
are trained on a varied dataset of
thermodynamic characteristics. Extensive
testing and validation proved that the system
could adapt to different operating situations
and provide higher diagnostic accuracy. This
study highlights the possibilities of machine
learning and the internet of things (IoT) to
revolutionize engine problem detection by
providing a dependable, efficient, and scalable
solution for contemporary industrial and
automotive uses. The research results help
improve engine management systems'
operational dependability and safety by
paving the way for predictive maintenance
solutions