IOT Based Predictive Maintenance Using LSTM RNN Estimator

Predictive maintenance is a smart solution for many industrial and commercial plants; it enables users to fix their devices before they fail. It is based on a mathematical model that predicts the failure time of the connected devices. In order to optimize the mathematical model large amount of data is required that increases the accuracy of prediction. Internet of Things (IoT) can provide these models with the required large data. Machine learning techniques have been used to produce increasingly effective solutions to predict the remaining useful life (RUL) of devices accurately. In this research a mathematical model is devised for general device that is connected to a central processing unit via IoT. The IoT provides a means of communication for the devices. It collects a huge amount of data from various sites. The large amount of collected data makes the prediction much closer to the true values. Two types of Neural Nets is used, Vanilla-RNN and LSTM-RNN, both showed good prediction results for a simple example of light bulbs. LSTM-RNN demonstrates better prediction results, where we recommend for critical devices. Other devices can use Vanilla-NN for its simplicity.