LSTM-Based Predictive Maintenance Framework for Turbofan Engines Using C-MAPSS Data

Abstract

In the modern industrial world, predictive maintenance is an emerging strategy to address the challenges of unexpected equipment failures, high maintenance costs, and production downtime. However, existing studies often focus primarily on prediction accuracy without providing an integrated framework that combines IoT-enabled sensor data with advanced deep learning models for Remaining Useful Life (RUL) estimation. To address this gap, this study proposes an end-to-end LSTM-based predictive maintenance framework for turbofan engines using the C-MAPSS dataset. The proposed approach leverages IoT-generated sensor measurements and Long Short-Term Memory (LSTM) networks to capture temporal degradation patterns and predict engine RUL. The performance of the LSTM model is evaluated against Gated Recurrent Unit (GRU) and traditional Recurrent Neural Network (RNN) models using Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE). Experimental results demonstrate that the proposed LSTM framework achieves superior predictive performance, attaining an MAE of 27.11 and an RMSE of 35.12, outperforming the benchmark recurrent architectures. These results highlight the effectiveness of LSTM networks in modeling complex sequential sensor data and detecting early signs of equipment degradation. The integration of IoT-enabled monitoring with deep learning-based RUL prediction provides a practical solution for real-time predictive maintenance. This approach supports timely decision-making by continuously analyzing sensor data and anticipating equipment failures before they occur.