Industry 4.0-Enabled Smart Monitoring and Fault Detection with Predictive Maintenance for Industrial Equipment

Authors:

V.B.K.P.D.Naidu¹, Botta Jayanth², Garikina Lakshmi Prasanna³, Nallabati Pavan Kumar⁴, Kandimalla Anitha Choudary⁵

¹Assistant Professor, Dept. of Electronics and Communication Engineering, Sanketika Vidya Parishad Engineering College, Visakhapatnam, Andhra Pradesh, India

^2,3,4,5B.Tech Final Semester, Dept. of Electronics and Communication Engineering, Sanketika Vidya Parishad Engineering College, Visakhapatnam, Andhra Pradesh, India

Abstract — We built this system because the problem it solves is real. Small manufacturing units in India lose days of production every year to motor failures that could have been caught days or even weeks earlier — if only they had a way to monitor what was happening inside the machine. The advanced systems that large industries use for exactly this purpose cost lakhs of rupees and need a dedicated team to run. That is simply not an option for most small businesses. So we asked a simple question: can we build something that actually works, costs under ₹5,000, and can be set up by someone with basic electronics knowledge? The answer turned out to be yes. This paper describes that system. An Arduino Uno reads vibration, temperature, supply voltage, and load current from a running motor every second. A NodeMCU ESP8266 sends this data to the ThinkSpeak IIoT cloud every 15 seconds over Wi-Fi. A MATLAB script running on ThinkSpeak then does something most low-cost systems cannot — it computes a live Health Index that tells you how healthy the equipment actually is right now, and it estimates how many minutes of useful life remain before a failure is likely, based on the temperature degradation trend. When the Health Index drops below 50%, the system sends an automatic email alert so the maintenance team can act before the damage is done. We tested the system against three different fault types — undervoltage, overheating, and abnormal vibration — and it caught every single one within 3 seconds at the edge, with cloud alerts reaching us within 60 seconds. We also solved a problem we ran into during development: when the NodeMCU and the MATLAB script were both trying to write to the same ThinkSpeak channel, about 1 in 5 writes was getting rejected. Splitting them into two separate channels fixed this entirely.

Keywords — Industry 4.0, Predictive Maintenance, IIoT, Health Index, Remaining Useful Life, ThinkSpeak, Arduino, NodeMCU, Fault Detection, MSME, MATLAB, Edge Computing, Condition Monitoring.