International Journal of Scientific Research and Engineering Development

This project combines Deep Learning (DL) for accurate prediction with SimPy for realistic simulation to address predictive maintenance in industrial IoT. Our main goal was to develop a reliable system by first building a highly accurate model to predict robot failures and then testing its performance under real-world conditions. We created a Deep Neural Network (DNN) that reached 1.0000 accuracy in classifying robot status. Next, we used this model in a SimPy environment to simulate maintenance queue management. The simulation showed that even with perfect prediction, the system still faced bottlenecks, leading to an average robot waiting time of 9.47 time units. This result shows that the main challenge is not prediction, but resource management.