RUL is very important for mechanical systems, in order to predict the remaining time to failure. This is essential for safety, maintenance cost and loss of time. This would be more pronounced in turbojets in aero planes and the like. This research presents a methodology to predict the residual lifetime and technical defect of mechanical components. The present model is applied, as a case study, to data supplied by NASA of a real turbojet engine (Williams FJX-2), in order to validate the model. The dataset was divided into dependent and independent variables. Data visualization and feature selection are used to make the model more accurate by utilizing NumPy, matplotlib, Pands and Seaborn libraries via Python Programming. The correlation between variables is used to build up the model. The present model is a linear regression one. The linear regression model simplifies calculations, and most importantly, linear equations make it easy-to understand interpretation on a modular level. Though, simple but provides accurate results. To overcome the increase in the root mean square values of results and to increase the model accuracy, the linear regression is coupled with Convolutional Neural Networks (CNN); this merge improved the results greatly.  The model results indicated very good agreement with real data of the turbojet engine, which gives confidence in the present model. The mathematical solution is capable to estimate the residual time for any mechanical component and not limited to the current case study.
 
Special Issue of AEIC 2024 (Mechanical & Chemical and Material Engineering  Session)