by Yash Bhardwaj
The aim of the project is to deploy a Prognostics(predicting time of failure) model which not only takes into account the statistical methods of predicting the lifetime and optimum maintenance time frame of machinery, but also takes into account real-time operation history from thousands of cycles of operation. Predictions made using a trained Neural Network on sensor data are known to be more accurate, with a focus placed on raw events data allowing for reliable estimates in tandem with Probabilistic approaches.
The final model will be a hybrid of raw sensor-based trained Neural network and probabilistic models that have been conventionally used, thereby allowing for an interaction between a number of factors that are otherwise neglected.