Abstract:
As many industries transition to Industry 4.0, machine health monitoring is the key focus
right now. Industry 4.0 is a technological revolution involving the Internet of Things
(IoT) and artificial intelligence. This industrial approach made it possible to collect huge
amount of machinery operation and execution data to automate detection of faults,
methods to reduce downtime, boost component usage, and extend their remaining usable
life. It is also because of evolution in computerized control and communication networks,
all of which are example of information approaches.
Machine health monitoring previous methods were not capable of amending real time
data which was obtained from devices or equipment installed in industry so mass scale
production requires automation for maximum output. This monitoring approach can help
increase equipment efficiency, lower energy usage, eliminate unplanned downtime, and
prolong machine existence. Predictive Maintenance (PdM) is unavoidable in Industry 4.0
for long-term smart manufacturing using machine learning (ML).
In this thesis, we proposed and developed a product that analyses state condition of
machine using machine learning and IoT technology. This device is made up of small
electronics components MCUs as ESP32 and STM32 that captures data from current
sensor, temperature sensor and vibration sensor send it to cloud platform which is further
processed using python to implement unsupervised machine learning algorithms on data
from air-conditioning system. After successful approach, results are displayed on as
operating on and off state of machine, defines the health, shows daily and cumulative daily
usage and lastly displays if any abnormalities.
